- pipeline/dump_bird.mjs: imports web/js/world/magpie.js UNMODIFIED (bare 'three' resolved to the
repo's vendored build through a node resolve hook) and dumps birdGeometry() — so the A/B is
against Lane B's actual bird, not a port. MEASURED 182 tris / 216 verts. --sim walks a player
past a territory on magpie.js's own clock: drawn 54.1% of frames, and OF THOSE perched 80.5% /
swooping 7.7% / returning 11.8%.
- pipeline/bird_to_glb.py: wraps the dump as a GLB in E's frame (head +Z, +Y up) so the identical
render_views.py rig shoots both; --fold applies the perch pose (x x 0.42). glb_stat re-measures 182.
- pipeline/render_views.py --noemit: strips emission from BOTH candidates. normalize.py's
emissiveFactor 0.28 is albedo-MODULATED; a vertex-coloured mesh cannot express that in glTF and
Blender writes a flat 0.28 that lifts a black bird to grey. Off both sides or the picture lies.
- pipeline/view_sheet.py --ab: the strip as a GRID (one line per candidate, stacked) plus two
linear motion-blur blocks at 25% and 100% of the bird's width, and a wrapped notes footer.
- docs/shots/laneE/r39_magpie_ab.png — the sheet Fable rules on.
Recommendation in LANE_E_NOTES; picture first.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A. DEPOT — closed, under John's R22 standing Lane-E authorization (the authority that
shipped sit.glb in R28 and the nine R36 sweep assets). Pushed via the passwordless
tailnet ingress: _published.json 53 -> 56, drift check clean. Then verified the way
this lane verifies — FETCHED BACK OFF THE PUBLIC DEPOT and hashed, not "the upload
said 200": magpie/paling_fence/water_tank all HTTP 200, byte counts exact, sha1
identical to local, 3/3. _r38_results.PENDING_PUBLISH.json renamed, build_manifest
re-run (GLBs 33 -> 36), validate_manifest 0 errors / 0 warnings.
B. THE MAGPIE TINT — done, not offered. The bird shipped as an all-black crow (R38 4d
traced the loss to trellis2_mlx's PBR bake). "Hand-tint the atlas" is not a paint job:
the bake is a smart-UV unwrap of ~100 islands and none of them announces itself as a
nape. So pipeline/tint_atlas.py defines the tint in the MESH's frame and projects it
through the UVs — rasterize every UV triangle to texel->(position,normal), evaluate
soft anatomical windows on the positions (white nape behind a black crown, shoulder
bar on the flanks only so the dorsal midline stays black, rump + tail base stopping
short of the black terminal band), paint keeping the bake's own luminance as shading,
then dilate 5 texels into the UV padding so no black fringe survives bilinear/mip.
3-D masks give one consistent soft edge across every island at once; a 2-D blur could
not (it would bleed island-to-island through the black background).
GEOMETRY UNTOUCHED byte for byte: still 894 tris / 1 mtl / 1 img / 512^2 WebP / no
Draco / 0.155x0.358x0.404 m. Painted luma median 0.84 (p90 0.95) vs the bake's 0.13.
VERIFIED IN THE GAME'S OWN RENDERER: the PUBLISHED file loaded off digalot.fyi through
the vendored GLTFLoader + three.js — 1 mesh, 894 tris, 512x512 texture decoded (which
also proves the PIL-written WebP decodes in a browser).
VERDICT: it reads. Pied and unmistakable at 64 and 48 px, still black-and-white at 32.
Honest caveat: from directly below-and-in-front it stays dark — correct, a real
magpie's bib and belly ARE black. Render: docs/shots/laneE/r38_magpie_tint.png.
Standing after B's 4514cb9: an UPGRADE PATH, not a dependency — B's 182-tri procedural
bird is 4.9x lighter and the right call. Nothing probes for the GLB and nothing should.
C. CORRECTIONS FOLDED INTO THE RECORD so the next asset round does not re-learn them:
skins are 9.7 s not 5.7 s (1024^2, 1.78x the pixels, per-pixel identical) - the reuse
list's "one line each" holds for 5 of 9 (extractGLB keeps only the FIRST material) -
normalize.py's collapse decimator floors at the shell count on TRELLIS output while
bake_lowpoly.py clears it every time - the 489 s outlier at 3.8x median means
throughput plans on a tail, not a mean - furniture.js:30's stale comment is fixed by B.
NEW, and the one worth keeping: an asset's tri budget is tris x THE INSTANCE COUNT THE
PLACEMENT RULE IMPLIES. paling_fence is a good 1,121-tri panel and a rejected asset —
905 instances = 1.01M tris, ~47x over. Ask the consuming lane for the instance count
BEFORE generating anything that tiles, edges or repeats per-lot.
New tools, all on-device and $0: tint_atlas.py, render_views.py (view sheet + --eevee +
roster-consistent --thumb), view_sheet.py (the distance strip no 256px thumbnail can make).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The town is inhabited: 23 towns finally differ, 181 back yards have a Hills Hoist with washing on it,
a magpie owns one street for six weeks of the year, and six point sources put a life off-screen.
MEASURED, walked, not bookmarked — two port-isolated no-store servers (:8472 = `git archive HEAD`,
:8471 = treatment, assets symlinked so E's mid-round wave is identical on both sides), 102 8-metre
spine steps at yaw 0, each frame fully streamed, through the shell's own composer:
worst draws 224 -> 225 (+1)
worst tris 89,502 -> 93,298 (+3,796 of ~75k spare)
sum draws over 102 frames 16,676 -> 16,663 (-13)
The citizen sim and the tram are hidden identically on both sides, and that IS the measurement: with
them live the same step read control 346 vs treatment 232, because peds spawn on wall-clock time.
`?yards=0` reproduces HEAD exactly (224 / 89,502 / 16,676). Isolated at a fixed pose: washing +1 draw
/ +512 tris, magpie +1 draw / +182 tris, every yard box prop +0 draws.
§1.2 THE CHARACTER VECTOR — +0 draws, +0 tris, classic byte-identical BY CONSTRUCTION.
`townCharacter(null)` returns the frozen v1 literals and a null tint keeps the SAME material cache key,
so classic gets the same object, not a copy of it: default-boot ground hashes identical to HEAD
mesh-for-mesh, `?classic=1` hashes unmoved from R37. There is no `if (classic)` in the ground path.
The sky feeds the BASE skyName, never `setSky` (a one-way latch that would have killed R32's dawn dome
on all 23 towns). The vacuous arm is per-town: `newtown_godverse` / `redhill_godverse` carry `state: ''`,
so a per-corpus arm would have dropped exactly those two silently. 21 distinct palettes across 23 towns.
`gravel` and `reddust` were not unwired but UNWIREABLE — their `use` values named no slot ground.js had;
making the slot addressable fixed it at the root, and both are selected now. E's three new grounds
landed mid-round and the upgrade was a name change in one table, which is what procedural-first is for.
§1.4 BACKYARD AUSTRALIA — inside boxMats, +0 draws. Houses inset off their boundaries so there is a
yard; 36 `yard` lots that had rendered as blank shopfront sheds since v1 are now fenced blocks with a
hoist, a shed and often a tank. Collision byte-identical (the lot rect is where the fence stands).
THE WASHING SHIPS, and I wrote the inverted weight rather than dropping it. The shipped top-weighted
mode moves a hem 0.10 cm calm / 0.67 cm gusty; `pegged` gives 4.5 cm / 31 cm — ~52x, and at the hem
instead of the pegs. The `canopy` string is byte-for-byte unchanged, so the gum-tree covenant argument
is untouched. It is ONE TOWN-WIDE InstancedMesh (1 draw at any N, cap 256, proximity-first off the R37
chunk seam) rather than per-chunk, which would have been +1 x 31 live chunks = 4x the whole margin.
Zero on every real town (no hoists ⇒ no mesh) and zero at night.
§1.3 THE MAGPIE — +1 draw on screen, exactly 0 otherwise; 182 tris of 900. Territory scales with the
road graph: 6 on the 6.78 km synthetic town, 157 on bendigo's 188.6 km / 2,593 edges, nearest 120 m
from spawn. The latch sets, the day roll clears it, localStorage gains no key and the save never
mentions it. `swoopPos()` is exported and pure. The hat stays cut.
MY OWN SEASON MECHANISM WAS VACUOUS AND THE SWEEP CAUGHT IT: keyed on citySeed alone it produced a
CONSTANT, because the town key is deliberately off the plan — 24 of 24 towns were 214 days from
swooping season, and it would have gated green because `?magpie=1` forces it. Re-keyed on
(citySeed, townKey): `newtown_real` is in season on arrival on the default seed, marrickville is 8 days
out, and 12.9% of 1,000 seeds put the synthetic default in season against a design ratio of 11.5%.
§1.1 THE AUDIO LAYER — six new mixer layers, zero draws, zero tris, seeded anchors, distance falloff,
AND A StereoPannerNode PER LAYER, so direction is real (measured spread -0.71..+0.85) instead of a
limitation written into the notes. Day-of-week derived locally. All six anchored on the default boot;
3 of 6 report `no-anchor` on real towns rather than being faked onto shopfronts. The falsifiability
control — one manifest key renamed away, served from a third port — makes that source report
`no-entry`, 0 plays, never `audible`, with the other five unchanged and zero console errors.
TWO DEFECTS MY OWN GREEN NUMBERS COULD NOT SEE, both caught by a screenshot: the washing shipped at 256
instances / +1 draw / cap respected and was INVISIBLE behind its own back fence (per-instance peg height,
and a fence taller than the hoist), and the magpie perched on THIN AIR over the road. R37's lesson
relearned at full price. Fence 1.60 / hoist 2.15 measured against the object; territories now snap to
furniture.js's own streetlight rule.
LANE E'S PROPS, ADJUDICATED: hoist + aerial permanently primitive (E's spike killed them, and primitive
is also the +0-draw mechanism). `paling_fence` at 1,121 tris x 905 runs = 1.01M triangles — no. The
`magpie` GLB is HELD FOR THE TINT: it reads as a crow, and a magpie seen for 1.2 s in peripheral vision
IS its white bar. My 182-tri bird already carries it, at a fifth of the triangles and no fetch.
furniture.js's "novelty_record is 26.5k tris" was stale by six epochs — corrected to 8,000.
Zero fetch delta: `?noassets=1` fetches the same 23 assets, same list, on HEAD and on R38.
0 console errors on 10 boots (default, classic, yards=0, washing=0&magpie=0&ambient=0, noassets,
magpie=1, darwin, hobart, bendigo, the patched-manifest control) beyond the known PointerLock ones.
Shots: docs/shots/laneB_r38/. Detail + every number: LANE_B_NOTES §38.
A. PIPELINE FIX: gen_props.py trellis_mac -> trellis2_mlx, contract read from its
manifest.json. The two operators are NOT drop-in on output — trellis2_mlx writes
BOTH candidate.glb (raw) and candidate_pbr.glb (baked), so the old
sorted(rglob('*.glb'))[0] would have silently shipped untextured meshes. Added
_pick_glb() (result.json first) + the bg_remove_local cutout stage, fail-soft.
MEASURED: median 128.0s vs trellis_mac's 245.4s p50 = 1.92x. 5/5 first attempt,
0 retries against the 4-5% budget. One 489s tail outlier, no error — named.
B. AUDIO (1.1) — six PACK entries, +0.86MB, ZERO draws/tris, all originalby
construction: dog-bark-fence, mower-distant, radio-through-window, sprinkler,
roller-door, magpie-carol. Seeded via crc32(name); beds fold seamlessly
(wrap-jump <=0.18x internal step). Verified by spectrogram + band-energy, and
one defect found by LOOKING: the sprinkler's spray masked its own ticks and read
as a hose — rebalanced and re-rendered. Distance is spent in the SPECTRUM, not
the level, because encode()'s loudnorm undoes a quiet render.
C. GROUNDS (1.2) — gen_skins.py had NO ground template; added one with its own
outdoor style prefix (the house STYLE says "warm fluorescent", an interior cue).
4 new 512^2 square tiles. Review's 5.7s is 9.7s here — same per-pixel, mine are
1.78x the pixels. Found: gravel/reddust are selected by nothing partly because
their `use` values name no slot ground.js:177 has.
D. MESHES — 3 shipped (magpie 894 tris, paling_fence 1121, water_tank 500), all
GLB-law clean. 2 FAILED and were DELETED: hills_hoist + tv_aerial. Thin wire
returns as thousands of disconnected fragments, so collapse floors at 250k/89k
AND a remesh coarse enough for 600 tris erases the pole. Lane B's boxMats
primitives cost +0 draws and are strictly better. The magpie's geometry passes
but TRELLIS crushed its white nape/wing to black — traced to the operator's bake
(the cutout and our decimation are both exonerated); flagged, not hidden.
E. REUSE — all 9 idle assets re-measured; the review's "one line each" holds for
only 5. furniture.js's extractGLB keeps only the FIRST material, so park_bench/
coffee_table/box_crate render in material #1 and bench_wood (0 materials) falls
to grey. Nothing was regenerated that already existed.
The 3 GLBs are staged, NOT published: a depot push is outward-facing and had no
in-session authorization. Held at _r38_results.PENDING_PUBLISH.json; publish +
rename + rebuild lands them. Gate green: validate_manifest 0 errors / 0 warnings.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
§0.1 — ground.js adopts Lane A's corridor law. The road quad narrows from the whole
corridor to the carriageway and the EXISTING merged footGeos class extends inward to
meet it, consuming `vergeBand(e)[0]` (not `roadWidth/2` — per A's mid-round correction,
so the geometry and A's corrected gate read the same number).
MEASURED, fresh boots, port-isolated no-store servers (:8178 = HEAD control, :8177 =
treatment), 2560x1440:
synthetic ground 5 draws / 224 tris -> 5 draws / 216 tris
walked spine, 102 steps @ 8 m: worst 276 draws BOTH; sum of draws
20,081 BOTH (identical at every step); 8 bookmarks identical draws
katoomba_real ground 4 draws / 9,932 tris -> 4 draws / 9,644 tris
walk 120 steps: worst 100 draws BOTH; sum 7,703 BOTH
+0 draws is proven, not asserted. Tris go DOWN (-8 / -288): the charter's "~2k tris" was
an estimate against a class that already existed. Honest number, stated loudly.
Surfaces verified by RAYCAST, not by reading the code (offset from centreline, topmost hit):
main@28 road 0-4.75 | kerb 5 | footpath 5.25-17.5 (band [5,14])
main@24 road 0-5 | footpath 5.25-15.25 (band [5,12], real)
side@12 road 0-3 | footpath 3.25-9.5 (band [3,6] — 72% of the corpus)
lane@8 road 0-4 | footpath 4.25-9 DEGENERATE band [4,4] => pre-ruling geometry
lane@4 road 0-1.75 | kerb 2 | footpath 2.25-5.25 (synthetic, also unchanged)
arcade footpath 0-6.5, NO road quad, NO kerbs (1,102 corpus edges)
RESIDENTIAL side@14 road 0-2.75 | kerb 3 | GRASS 3.25-6.75 | footpath 7-10.5
The lane case branches on `outer <= inner`, never on a band literal.
Posters: 14 of 14 stood inside painted bitumen before; 0 of 14 after. Not one poster moved
— the ground did.
ONE DEFECT FOUND AND FIXED IN THE ROUND. The first cut painted every intersection shut:
a 12.5 m footpath crosses the perpendicular CARRIAGEWAY where a 3.5 m one only ever
clipped a far corner. Screenshotted, rejected, and fixed by putting the carriageway on
top (ROAD_Y 0 -> 0.04 under the flag): roads run through junctions, footpaths stop at the
kerb they meet. Verified on the real graph's worst 4.3-degree junction.
?verge=0 ships permanently — the gate's falsifiability control. Proven EXACT, not merely
similar: ground vertex+uv hashes identical to the HEAD build (c4c72181 / c111d968 /
1a80ebcd / 12c1241b / fd0d3211), and the same detector goes RED on the default boot.
CLASSIC IS FENCED, NOT AMENDED (the R31 dig-flip pattern). ?classic=1 forces verge off, so
the covenanted town's render does not move on a render-side ruling. Proven twice: identical
ground geometry hashes AND identical ground-only rendered pixels at three poses
(79df268c / 91564001 / a09d1e52), while the default boot differs at all three. If Fable
wants classic to take the footpath too, that is an amendment ruling and a one-token change.
§0.4 — the chunk lifecycle seam. The producer has fired since R3 and Lane D's consumer has
existed since R3, but nothing could connect them: createChunkManager never handed `ctx`
back, so a consumer holding only PROCITY.chunks had nowhere to attach (measured on HEAD:
`!!chunks.ctx === false`). Now `chunks.onChunkBuilt(fn) -> unsubscribe()` for many
consumers plus `chunks.ctx` for the original single slot. Proven live: 136 build / 144
dispose events over a 40-step walk; payload {cx,cz,buildings,furniture,data} with
key === `${cx},${cz}`; a second consumer does not clobber the first; the ctx slot fires;
unsubscribe silences; a THROWING subscriber does not take the streamer down.
ZERO behaviour change when nobody subscribes: a 102-step walk recording
[draws, tris, chunks] per step hashes ceb1ec66 on BOTH builds.
Files: web/js/world/ground.js, web/js/world/chunks.js, docs/shots/laneB_r37_kerb/*,
docs/LANES/LANE_B_NOTES.md, B-progress.md. sim.js untouched (Lane D's).
The v7.0 charter's money shot, driven through the REAL loop on the pinned seed: arrive hunting
(the splash mission) → the strip → the shop (rumor overhead) → THE GEM (THE VERANDAHS 'Tinny',
asking $19 against ★ RARE guide $60+ — with onboarding beat ③ visible below, walking its talk)
→ the crate → dawn day 2: '⚖ SOLD Tinny — hammer $89, $78 after fees'. Bought $19, hammered
$89: the loop, told in six frames. docs/shots/v7_tour/ + sequence README. What remains for the
v7.0 tag is deliberately NOT mine: docs freeze + John's playtest (hands before tags).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
qa.sh --strict --matrix: 7 passed · 0 failed · 0 warn · 0 skipped, MATRIX 10 towns green. New
STRICT smoke_first_five: the fresh game hunts a roster artist and the splash names hunt + cash
· the beats walked the REAL paths (② at the crate, ③ on the buy, the send-off on the ⚖) · a
returning save gets no tutorial · a second fresh browser hunts the SAME artist (seeded, not
random) · ?game=0 shows nothing · 0 console errors. THE CROSS-ROUND CATCH: the seeded want
moved a fresh-boot invariant (wants 1, not 0) and the R33 gate went red on the FULL run while
green solo — amended to a baseline snapshot, lesson filed in §35 (a feature that moves a boot
invariant must sweep every gate that assumed it; full runs exist for exactly this). Money shot:
docs/shots/v7_beta/first_five_splash.png. Remaining for v7.0: the tour + docs freeze + THE TAG
— which waits for John's hands (the alpha precedent).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
qa.sh --strict --matrix: 7 passed · 0 failed · 0 warn · 0 skipped, MATRIX 10 towns green. New
STRICT smoke_godbay: consigned through the real crate DOM · ONE-HAMMER determinism (a frozen
pre-sleep export slept twice lands byte-identical — the seeded gavel, measured not trusted) ·
honest bounded resolution (net $19 ≤ 2×$24×0.88) · the wake toast carried the morning paper ·
the ride to melbourne hammered a listing on arrival · ?game=0 has no crate to list from · 0
console errors. Money shots in docs/shots/v7_beta/ (the dawn paper + the ⚖ crate). The full
v7 loop now closes end to end: dig the gem → keeper cash now or the gavel tomorrow → the
wantlist points at another town → the ride costs a day and hammers your listings on arrival.
Remaining for the v7.0 tag: THE FIRST FIVE MINUTES (onboarding), the tour, docs freeze — and
John's hands on the whole thing.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
qa.sh --strict --matrix on the final tree: 7 passed · 0 failed · 0 warn · 0 skipped, MATRIX 10
towns green. New STRICT smoke_wantlist_travel: ☆→★ through the real card (the save carried
{THE VERANDAHS — 'Servo at Midnight'}) · a keeper muttered the hunted artist within 2 shops
('saw one in Castlemaine just last week') · buying the wanted sleeve cleared the hunt · the
synthetic→melbourne ride cost EXACTLY one day with $179 and the find riding along · the
same-town reboot was free (the control discriminates) · the ride home charged again (day 3) ·
?game=0 built zero want DOM and muttered nothing · 0 console errors. Matrix note: town_matrix
shares one browser across towns so travel drifts the day as it walks — safe because
determinism double-boots the SAME town (free by the same-town rule); reasoned before landing.
Money shots: docs/shots/v7_beta/rumor_and_wantlist.png (the keeper's mutter over the toy
shop's 70s wallpaper + THE WANTLIST open in the crate) + want_card.png. README manual row.
v7.0-beta systems are now ALL BUILT — what remains of v7: Esc-in-person (John's hands), then
the v7.0 close (GodBay offline auctions, THE FIRST FIVE MINUTES, the tour, the tag).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
qa.sh --strict --matrix on the final tree: 7 passed · 0 failed · 0 warn · 0 skipped, MATRIX 10
towns green. New STRICT gate smoke_bible: band law over EVERY bin off the digOffers surface
(80 offers/5 bins), the buy-everything ledger runs on THE GEM'S BIN so the finder-margin leg is
non-vacuous (bin1: 2 gems, grail asking $19 vs keeper $30; ledger closes to the cent — Σask
$333 → Σoffer $135, gems alone +$21, every non-gem strictly negative), scarcity measured
(emptied bin 0/16 on re-open AND across a full reload; 16 pulls in the save) with the ?game=0
negative control re-opening 16/16 (the detector discriminates). sell_routing amended from the
alpha inequality (credited < paid) to the exact formula (credited == min(basis−1, basis//2),
basis = bandLow-or-pricePaid) — the fence moved from the receipt to the band, which is what
charter law #2 always said; it passed on a banded standard item (+$4 on basis $8). _at_bin
grows an idx param (back-compatible). Money shots in docs/shots/v7_beta/ (the guide card + the
crate at the counter). LANE_F_NOTES §32 (incl. the JING5 fresh-checkout trap: gitignored
web/assets/models synced from m3ultra — also un-warned the validator's 15 cross-pack checks).
README manual: the guide/gem/scarcity rows. Remaining beta charter queued in the round doc:
wantlist + rumors · travel costs a day · Esc-in-person · then GodBay.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
qa.sh --strict --matrix: 7 passed / 0 failed / 0 warn. selfcheck ALL GREEN 156,352/156,352 ·
MATRIX GREEN 10 towns x 7 gates · scaffold_check GREEN (F's re-pins took) · flags harness green,
carrying the kill-the-server gate and the beta crate gate.
F RE-PINNED ITS OWN FILES per A's handoff (A doesn't edit other lanes' files, correctly):
GOLDEN_HASH 0x3fa36874 -> 0x5f76e76 · GIG_GOLDEN 0xb1d48ea1 -> 0xec7a2d39 · melbourne · katoomba
· the captions. AND A CAUGHT A LIVE HAZARD IN F'S GATE: hash_ok carried `or ('3fa36874' in
r.stdout)` — a bare-string OR-clause, dead the moment the covenant moved and worse than dead,
because it would go green off a SENTENCE ABOUT the old covenant appearing in stdout. Deleted.
The check derives from GOLDEN_HASH or it doesn't exist.
JOHN'S RULING IS THE EPOCH'S BEST LINE: the covenant is anti-drift machinery, not a defect
preservative. 0x3fa36874 did its job for five epochs — it never drifted. It just was never right,
and no hash can catch what was wrong in the first commit. A golden proves NOTHING CHANGED; it
cannot prove anything is CORRECT. Every gate we own was green over a world of back walls, and the
only thing that ever saw it was a human standing in the street looking at a wall.
THE FRONTAGE FRAME TOOK FIVE ATTEMPTS, FOUR OF THEM F'S FAULT:
1-3. cluster pose (the town's heart isn't this shop) / lot-anchored facing along the edge /
lot-normal facing — three photographs of empty road. Also never winnable: the shops faced
backwards the whole time. F was posing a camera at a wall and tuning the geometry.
4. B's helper — and F THREW THE RESULT AWAY. poseForShopFront RETURNS a pose; it does not
apply one. F called it, error-checked it, and never touched the camera, so the frame was
the spawn paddock — and F's error-check PASSED, because the call succeeded perfectly. The
vacuous check one last time, in the last frame of the epoch.
5. Applied it -> MONSTER ROBOT PARTY. Sign, door, awning, a ped walking in off Musgrave Road.
The measurement that would have found the facade bug in five minutes — compare the two
conventions against each other — is the one nobody ran for nine rounds, F included, right up to
the vacuous 72/72 F nearly filed as "B is wrong".
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
THE KILL-THE-SERVER GATE, permanent and green (charter law #1 made falsifiable):
ARM 1 server-up · gone[3] → 117 survivors · items[] 120, identity AND ORDER intact ·
COLLATERAL 0 · `gone` carries the atlas id verbatim (R26 fence)
ARM 2 killed mid-session · enter 6ms vs control 6ms (Δ+0) · worst frame 204ms vs control
202ms (Δ+2) · room complete from files · ZERO game errors · EXACTLY 1
refused connection
ARM 3 server-down boot · pure tier 1, gone[0] · ZERO game errors · exactly 1 refusal
SANDBOX · POST/PUT/PATCH/DELETE all 501 · a mint shop 404s
The subject exists only because G built `make-fixture`: production `gone` is EMPTY (0 sales
since 2026-07-01), so this gate would have passed vacuously forever on live data.
CORRECTION 1 — "ZERO console errors on server death" is UNACHIEVABLE, and F measured why. A
fetch to a dead host logs net::ERR_CONNECTION_REFUSED at the BROWSER level; no .catch() sees it.
To log zero the reader would have to know the server is dead WITHOUT ASKING. One refusal is the
irreducible cost of discovering death. F's replacement is strictly stronger: ZERO errors from
game code, and EXACTLY ONE refusal — which PROVES the circuit breaker. Without the breaker it is
one error per shop entry forever, and "zero" could never tell a working breaker from a broken
one because both fail it. F is the beneficiary of this change, so it is filed loudly in §27 for
Fable to ratify or overrule, not quietly relaxed.
CORRECTION 2 — F's frame assertion was weak: `< 250ms` is a number with no control. ARM 2 first
reported worst-frame 201ms and PASSED. But entering a shop rebuilds an interior, which costs the
same with or without a server — so an absolute threshold measures Lane C's room build, not the
network. The arm now walks the same shop with the breaker tripped and asserts the DELTA: 202ms
control, 204ms killed, Δ+2ms. The death cost nothing the room didn't. F told A the same thing one
round ago; a number without a control is not a measurement.
THE TOUR — 3 of 4 frames, and the 4th is FILED NOT SHIPPED. The real crate (the money shot: its
own 120 records, price stickers, the keeper), a mint crate (captioned MINT — plausible, not
real), the classic covenant. 0 page errors. The FRONTAGE frame does not land after three
attempts and F will not ship a frame that doesn't show what its caption claims: the cluster
primitive answers "where is this town's heart", which is not the question a frontage frame asks
— Red Hill's heart is 2 shops/60m and does not contain Monster Robot. Next: ask B, who owns the
frontage and shipped its mark this round. Re-deriving B's anchor in a tour script is how you get
three bad frames.
Also: the interior beauty shot came back with the HUD across it — R21's lesson, fixed on the
street frames only. Now every DOM overlay is hidden, not a guessed selector list.
NOT TAGGED. qa.sh --strict --matrix and the frontage frame remain.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The blocking gate is CLOSED, and the epoch closes green.
- #1b THE BLOCKING RE-MEASURE: F's R21 re-measure made the tri diet blocking; B diagnosed layer-by-layer,
the driver traced A -> E (A's own commit: "B's numbers say the fix is E's"), E fixed it, F re-measured:
katoomba venue_night 221,935 (+11%) -> 108,365 (-51%); bendigo 201k -> 124,682; fremantle -> 88,397.
Worst 124,682 — the 200k stress edge is closed. The measure-first loop worked end to end.
- #3 DBG.enterShop(selector) (Lane G's ask): id · registry type · exact name · name substring, resolved
most-specific-first so an id-shaped string can't be shadowed. One-liner QA on any of 23 towns.
- #4 The release tour -> docs/shots/release_v4/: the Fremantle hero · the Fitzroy mecca · the Toowoomba thin
tail · the synthetic control · the ?classic=1 covenant. Every real-town frame poses through F's R21
cluster primitive; B's R22 bookmarks anchor on the same retail heart, so tour and bookmarks agree by
construction. HUD hidden throughout (F's R21 ruling corollary).
The primitive earned itself twice: the first cut posed via street_noon and the MECCA CAME OUT AS BARE
BITUMEN — the same trap as R21. Re-posed through the cluster, Fitzroy reads.
- HONEST FINDING (filed v4.x): the brief asked for the thin tail as "proof the floor holds". It does not,
and the frame says so — Toowoomba reads as an empty road. The floor is a RATIO, not a count: 0.008
shops/edge vs the heroes' 0.06-0.08. Shipped labelled honestly rather than swapped for a flattering town.
- #6 Docs: README v4 section (the towns ARE the product), LANE_F_NOTES §22, F-progress.
qa.sh --strict --matrix GREEN 7/7; MATRIX GREEN — 10 towns × 7 gates (the budget gate green everywhere for
the first time); selfcheck 161,300/161,300 (A's 36 goldens + the 23-town pack).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
All five lanes in (A's venue cluster bias, B's selector-from-index, C's density door-soak, D's relocation
verify PASSES, E's 22-town pack non-gating); F gated and tagged.
- town_matrix.py: a `tram` dimension — B's R20 shop-adjacency ruling verified EXACTLY (newtown runs 16 stops
after B's phantom-stop fix, castlemaine 10; katoomba/fremantle/bendigo FENCED at 3/2/1 shops fronted, all
below the 5-shop floor). 10 towns x 7 gates: boot/determ/district/roads/tram/noassets green on all ten.
- index.html: expose window.PROCITY.tram — B built `routeInfo` FOR F's smoke but nothing published it; F owns
the shell, so F wired the seam.
- RECORDED (ledger #6b — the instruction was "record the new number"): the venue_night re-measure. A's cap
WORKED (posters 438->111) but E's density widening ate it — katoomba 200,965 (+965, R19) -> 221,935
(+21,935, +11%); bendigo 201k (+0.5%); fremantle/newtown/castlemaine in budget. Draws fine everywhere
(151/300 — half budget); tris-only, at the worst night stress view. Filed as R22's first item, where the
tri-diet backlog (instrument LOD, tram tri diet) already lives. katoomba is the outlier: 966 edges, 72 shops.
- Ruling (ledger #6d): the town selector stays ALWAYS-IN-HUD, not ?dbg-gated — THE REAL MAP's product IS the
towns. Corollary (F's own gap): money shots hide the HUD; the R19 alpha shot had the selector in-frame
because F's harness screenshotted the full viewport.
- Beta money shot: Fremantle's real street at golden hour — a corridor of shopfronts, footpaths, a crossing,
gum trees, and a genuinely busy crowd (the density widening, visible). 126d/167k tris, HUD hidden.
- docs: LANE_F_NOTES §20 (the honest why-R20-had-no-tag record) + §21, README beta paragraph, F-progress.
qa.sh --strict GREEN; selfcheck 154141/154141 (A's re-pinned goldens hold).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A resolved the R18 findings (fragmentation cull/join, poster clamp, goldens pinned) + D verified; F gated.
- town_matrix.py: a `roads` dimension — a schema-v2 cache (roads[]) must build the REAL street graph
(norm.mode==='roads', read via the plan_osm opts.report sink); synthetic/fixtures stay marched/synthetic.
10 towns x 6 gates: real roads green everywhere; classic/default proven untouched (roads ride the cache
schema, not a flag — 0x3fa36874 frozen).
- docs/shots/v4_alpha/katoomba_mainstreet.png: the first real-street money shot — Katoomba's actual main
street (asphalt + lane markings, footpaths, a pedestrian crossing, a walking crowd, gig posters, B's town
selector in-frame), PROCITY's look on real bones, 121k tris midday.
- docs: LANE_F_NOTES §18 (why F correctly HELD in R18 — the gate would have failed by design) + §19 (the
close), README real-roads note, F-progress.
qa.sh --strict --soak GREEN 7/7. One filed NON-blocking finding: katoomba_real venue_night 200,965 tris
(+965 / 0.5% over the soft ceiling, poster-count-scales-with-real-edges driven; midday play is 121k) — a
v4.0-beta hardening item (A caps posters per real town), on the opt-in --matrix warn-gate, not a play view.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The closing half of F's bookend. All gates green on the flipped tree (qa.sh --strict --soak: classic
regression + default-boot gate + smokes + no-giants incl. the seated drummer + the full-week soak).
- tour_shots.py -> docs/shots/release_v3_1/: the money shots on the DEFAULT boot (no per-shot flag juggling
now the flip covers it) + a ?classic=1 before/after ("the town before v3"). 7 shots, settled, 0 page errors.
- week_soak.py: re-pointed to the default boot; the leak check now treats an at/below-baseline OR non-climbing
trail as leak-free (a decreasing tex trail is GC/atlas-settle, not a leak — it was false-failing at warn-level).
- docs: README v3.1 flags table (defaults on + ?classic=1), CITY_SPEC v3.1 amendment (the rewritten prime
law + classic covenant, A's deferred edit), LANE_F_NOTES §16, F-progress R16.
Numbers (seed 20261990): default @ venue block 124 draws / 57k tris (<=300/200k); ?classic 161 draws / 24k
tris (the v2 baseline); seated drummer 1.32 m stature; week soak 42 cycles leak-free (net +0/-6).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
tools/qa/tour_shots.py: one deterministic run capturing the release set into docs/shots/release_v3/
(+ contact.html), grouped by the flags each shot needs, all seed 20261990:
- venue_night, district_posters, queue_night (John's HERO — now with A's flipped frontage poster VISIBLE
by the door), gig_interior (the 4-piece + real instruments + a roster crowd — 2 punters carried through),
rain_verandah, tram_stop
- settled discipline (ledger #5): every shot measured after the async instrument GLBs / poster textures /
chunk stream land (shoot-twice), from a clean per-group load; the queue is driven LIVE (explicit admits,
not rAF auto-drain which automation throttles)
The hero confirms ledger #1: the frontage poster reads from the street for the first time. Interior settles
35k->182k tris (draws 56 <= 350 gate). All street shots <=300 draws / <=200k tris. 0 page errors.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Wires A–E + D's district together and fires it. Seed 20261990 → "Boolarra Heads": 3 venues, 2 playing
tonight (The Exchange Hotel pub gig-pubrock + an RSL gig-covers, one dark), Screaming Utes $10 on the door.
- gig_state.js → PER-VENUE latches off the shared procity:segment spine (one per venueShopId); all three
hard-won R12 laws kept. New runtime surface window.PROCITY.gigs.byVenue + onOf/openOf/gigOf/coverOf/
bandNameOf/paidOf(id) + markPaidOf/venueShopIds; the R12 single-venue alias kept for one round.
- interior_mode.js: crowd-walla rides the interior tone layer under the live band bed while `on`; an
applause sting fires once on the on->done edge. DELETE the GIG_BED bridge — debt #1, and it was a LIVE
BUG: E renamed pubrock-live->gig-pubrock, so the bridge (gig-pubrock->pubrock-live) pointed at a dead
key => the band would have mimed in silence. room.audio.gigKey (already the manifest key) feeds directly.
- index.html: per-venue cover charge (a night at two venues is two covers), concurrent per-venue setGig
surges, and the outdoor VenueQueue lifecycle (D shipped the class + left the street-side seam to F:
spawn at doors off B's queueZoneFor, drive each frame, dispose by done), exposed as queueCountOf(id).
- interior_scale_check.py: human-size assert now measures STATURE (crown-foot, platform-independent);
world-crown ceiling-clip assert kept as its own line (debt #6).
- flags_check.py smoke_gigs -> THE DISTRICT: band->4, per-venue loop (every gigKey resolves across 3
genres, crowd <= its own watchPoints, no-giants by stature), street budget at the busiest venue block,
the queue assert; plus the R12 determinism/cover/noassets/flags-off arms.
- CITY_SPEC: pin the window.PROCITY.gigs runtime surface. gig_shot.py -> r13 interior + venue_night shots.
qa.sh --strict GREEN 6/6, 0 warn. Interior 63 draws/35k tris; street @ venue block 121 draws/51k tris;
band+crowd stature 1.61-1.81 m under a 4 m ceiling; flags-off fingerprint 0x3fa36874 unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- theme.js: band_room (small/scuzzy tin shed) + rsl (carpet-and-flock club) venue recipes,
reusing the pub machinery via a new per-kind venueSpec (deck size/height + crowd cap).
- layout.js: parametrized placeStage/placeWatchPoints; 4th `drums` pose seated on the riser
(debt #3 — the band is now a 4-piece across all venues), stage.riserY + per-pose y.
- interiors.js: robust gig-<genreKey> resolution (gig.genreKey -> shop.genreKey -> kind
default) matching E's renamed manifest beds (debt #1); API doc for the 4-piece + riserY.
- LANE_C_PUB.md: 3-kind venue contract + pinned ry convention and the empirically-verified
keeper/browser back-to-front finding + source-normalise fix for Lane D (debt #4).
Verified fresh (seed 20261990 + 90-build soak): watch caps 8/8/12, gigKeys
gig-pubrock/gig-grunge/gig-covers, determinism 0, path/carve 0, gig-night worst RSL 237 draws
asset-free (16 rigs) / 77 GLB <=350, leak geo0/tex0. Shot: docs/shots/laneC/rsl_gig_r13.jpg.
Lane F: bump flags_check band==3 -> ==4 (tools/flags_check.py:674,794) — RED by design.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
F last: five landed lanes → one Friday night. Seed 20261990 → The Thornbury Hotel,
tonight SCREAMING UTES, $10 on the door. qa.sh --strict GREEN 6/6.
- gig_state.js (new, F-owned): quiet → doors (DUSK) → on (NIGHT) → done (DAWN) → quiet.
Rides lighting's procity:segment event (a poll only sees the segments it samples, and
missed the night roll entirely); `state` is a getter because B's audio reads it from its
own rAF while the player is inside a shop.
- ?gigs=1 seam: custom_bands.json fetched in the bootstrap (no fetch flag-off or ?noassets)
→ generatePlanFor(seed, src, {gigs:true, customBands}). 7 of John's 10 names fill the week.
- Cover charge on the R8 buy seam, zero new UI: paid debits once with a toast, re-entry the
same night is free, broke → polite knockback, free nights walk straight in.
- interior_mode: opts.gig → C's gigKey; D's GigCrew (band 3 + crowd 8, 3 dancing) at C's
stage/watch points; both disposed with the room (leak-free over 4 cycles).
- Cross-lane fix at the glue: C emits gigKey 'gig-pubrock', E shipped the bed as
'pubrock-live' — unresolved the pub plays to a silent room and nothing errors. Bridged
here; C/E/Fable owe CITY_SPEC a contract line. Details + 2 more findings in LANE_F_NOTES §12.
- smoke_gigs (16 gates): schedule determinism, band==3, crowd ≤ watchPoints, dance mix, bed
resolves, ≤350 draws (47 rig / 171 placeholder), no-giants by STATURE (the band stands on a
0.32m deck, so world-crown fails a correct band), cover ruling, ?noassets+?gigs.
Flags-off regression gains a v3 arm: no state machine, no plan.gigs, no pub, no venue geo.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
v3.0-alpha — Friday night at the pub. New web/js/citizens/band.js (GigCrew):
- Band trio on C's stage.bandPoses (deckY), distinct fleet peds (seeded per gig), each with a
primitive instrument on the chest (guitar/mic/bass; E's GLBs drop in via opts.instrumentFor).
- Crowd (<=8) at C's watchPoints on the gate-protected spawnRig path: dance:false = weight-shift
idle, dance:true = bounce/sway/step with seeded phase (no metronome).
- Deterministic from citySeed; placeholder crew under ?noassets.
sim.js: setGig(venueShopId, on) gig-night surge — patronage pulls peds from further (GIG_RANGE)
and ducks in harder (GIG_SURGE) at the venue; occupants drain at close per A's closing-time ruling.
Guarded by _gigVenue==null so gig-off is byte-identical (prime flag law; goldens unmoved).
Verified standalone (A withGigs + C buildInterior + D): band 3, crowd 8, dancing 3, all 11 figures
human-sized (1.60-1.83m under a 4.0m ceiling; R10 no-giants holds), band faces audience + crowd
faces stage, deterministic, leak-free (8 cycles, 0 delta). Crew ~40 draws. Surge: venue occupancy
1 -> peak 5, identity stable. Shot: docs/shots/laneD/r12_gig_night.jpg.
Flag for C/F: band.js flips C's watchPoints/bandPoses ry by pi (RY_FLIP) — the fleet mesh's visual
front is local +Z, not the -Z the comments assume. Remove RY_FLIP if C re-fronts the gig poses.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Verified Lane D's R10 buildFigure span-normalize fix: qa.sh --strict GREEN 6/6 incl. the new
no-giants scale gate (24 interior figs / 6 shops all crown 1.4-2.0m, under-ceiling). The 3.83m I
first measured live was a STALE ES-module cache in a day-old browser tab, not a regression
(served rigs.js + fresh-chromium gate both correct).
Re-shot the v2 tour on the fixed tree: browsers.png now shows human-sized figures (browser at the
rack + keeper at the counter). night_crowd.png re-framed to a readable clean-night "lit VIDEO REGAL
amid CLOSED town" (pulled out of the rain group; reuses night_neon's tighter dist-12 pose).
Honest finding: night_crowd has no visible street crowd and can't without new work. Lane D's night
patronage parks ducked-in patrons hidden INSIDE the shop (they are the interior browsers), and this
seed's video shop is a low-traffic corner (tested: 0 patronage / 0 visible at night). Flagged for D;
a visible outdoor night queue is a v3 mechanic. ~20fps in night captures is a headless software-GL
artifact (scene is light: 128 draws / 65k tris).
Full-default soak PASS (heap 54MB stable, no perf drift, 0 errors). Tour README updated with stable
v2 filenames. v2.0 annotated tag staged separately (push held for John — public release + the
night_crowd caveat on his flagged money shot).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
R9 blocker: interior rig figures were ~2x too tall (keeper crown 3.83m in a 3.4m room,
clipping the ceiling). Root cause in rigs.js buildFigure: it normalised scale by height/headY,
where headY is the head bone's world Y in BIND POSE — but the Mixamo skeleton origin is the hips
(~mid-body), so headY is head-above-hips (~half the standing height), not feet->crown. Every rig
came out ~1.83-2.0x too tall. (The head does NOT move when posed — bind crown 3.209m vs idle
3.205m — so it's a hips-relative measure, not an animation issue.)
Fix: normalise by the feet->crown span (headY - minY, both already computed). buildFigure(rig, h)
now lands the crown at exactly h above planted feet. Fixes keepers, browser rigs, near-tier street
rigs AND the impostor bake (all ride buildFigure). Verified: all 19 fleet rigs crown==1.75m; live
keeper 3.83m -> 1.82m; determinism + dispose-leak soak unchanged.
Regression guard (required): new tools/qa/interior_scale_check.py (Playwright) enters a sample of
shops, injects browser rigs at browse points, asserts every interior fig crown in [1.4,2.0]m AND
< room dims.H AND feet planted. Wired into tools/qa.sh --strict (run_gate => a giant fails qa).
Proven bidirectional: PASS on the fix (24 figs/6 shops), FAIL on a reintroduced giant. Full
qa.sh --strict: 6 passed, 0 failed.
Re-shot laneD/r10_browsers_fixed.jpg. Figures human-sized relative to doors/fittings: yes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- tram.js: 1 small bus, seeded out-and-back loop on the main-street spine,
door-dwell at each busShelterStops stop, ping-pong at ends. Deterministic,
exactly 2 draws, night headlights via lighting.isNight(). Flag-off identical.
Shell wiring (2 lines) documented for F in LANE_B_NOTES.
- dbg.js: 8 v2 tour bookmarks (16 total) incl. rain_verandah hero shot;
tram_stop now frames the nearest-origin (open) shelter stop.
- weather.js: cap rain gl_PointSize (clamp 2-22px) - near drops are thin
streaks now, not blobs. Rain smoke still GREEN.
- qa.sh --strict GREEN, golden hash frozen, flags-off byte-identical.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Follow a ped into a shop and find them browsing. D owns occupancy truth; C owns browse points;
F wires the handoff. Validated end-to-end in-shell (?stock=real). qa --strict GREEN; v1 untouched.
Occupancy (D1): patronage records who's inside which shop by shopId. New
citizens.occupancyOf(shopId) -> {count, occupants:[{seed,enteredAt,pedIndex}]} (pedIndex = the
exact ped who ducked in, for identity continuity). Cleaned on emerge/chunk-drop/stream-toggle.
Deterministic: 17 shops + occupants byte-equal across two runs (fixed a stale-occupancy leak on
stream toggle).
Browsers: keepers.js gained browse:true (faces the shelf via ry, no player-greet) + pedIndex +
seedKey. Each = a merged ped ~1 draw. 12 enter/exit cycles: count == min(occupancy,3) 12/12,
worst 61 draws <=350 (C's headroom holds), leak-free (0 GPU delta after warmup), disposed every
exit via keepers.disposeAll().
Consistency (D2): inside peds hidden on the street (0 rendered), re-emerge at the door.
F handoff verified by running F's exact interior_mode wiring (enter -> occupancyOf ->
browsePoints.slice(0,min(count,3)) -> keepers.spawn). index.html setShops door points need
shopId:s.id added. Closing-time handled by the dwell model (pending A's explicit ruling).
Evidence: docs/shots/laneD/r9_browsers_in_shop.jpg. Changed sim.js + keepers.js + test page.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Book/toy shelves get pull-and-buy over the SAME wallet; records keep the
deep dig. New stockpack.buildBuyableShelf(): a real shelf is merged,
per-item-ADDRESSABLE meshes — one per atlas (all covers drawn from one
seeded atlas → one draw/shelf), tagged noBatch so batch.js leaves them
intact. Aim a spine/box → pull card (title/author-or-maker/$price/band
from the pack index) → BUY debits the wallet, adds to inventory, and
collapseBuyItem() zero-areas the item's quad in place (removed from the
shelf, no geometry churn). Parody stock → no buy mesh → no card. Fail-
soft per pack; atlasPool() keeps each shelf single-atlas (one draw).
Test page: E interacts (bin→dig, shelf→card) + a #shelfBuy pull card;
shelfBuySoak() buys one item per book/toy seed. Validated: raycast→card
→BUY end-to-end (quad collapses to centre, cash/inventory update),
buy-soak 6/6 collapseOk leak geo0/tex0, ?stock=real draws record 41/
book 58/toy 57 ≤350, drawSweep GLB-off 168/on 344 unchanged, placement
determinism 0 fails, qa --strict GREEN 5/5. Shot: browse_buy_r9.jpg.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Shell seams (F owns index.html): A &town=<key> -> generatePlanFor(seed,'osm',{town}) (Katoomba);
B weather (createWeather on ?weather=1; PROCITY.weather={state,intensity}, {clear,0} off; update in
loop); C session wallet bound to the dig BUY (getCash/onBuy in interior_mode); D patronage door
points -> citizens.setShops + ?patronage=0 off-switch + setWeather each frame.
Harness: stock=real + weather=rain graduated warn -> STRICT. New warn smokes: buy-v0 (dig BUY ->
wallet, cash 191->179 bag 0->1), patronage (7 peds visit shop doors/~18s), book/toy stock (6/11
real atlas meshes). Determinism gate: pinned osm/katoomba golden 0x0f652510 keyed on (seed,plansrc,
town) + fixed a hex zero-pad bug the katoomba golden exposed (0x0f652510 vs unpadded 0xf652510).
No patronage baseline re-pin needed (nets fewer draws, within tolerance). qa.sh --strict GREEN 5/5.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
C1 — buy loop v0 (new wallet.js): runtime-only economy — seeded cash
(~$60–200), buy()/broke-gate/in-memory bag, onChange for a cash chip.
World gen untouched (purchases never re-enter the plan/seeded build).
Test page: #cashChip HUD + I-toggle inventory + dig BUY IT button
(pull → bag, cash down, bin depletes); digSoak buys ≤2/visit.
C2 — book spines + toy boxes wired to ?stock=real. stockpack.js keys
each pack to a slot (record→sleeve, book→spine, toy→box); stock.js
shelfLine builds real atlas-UV cover planes for book/toy shelves
(static, no dig), fail-soft per pack. E's packs live: record 350/6,
book 311/4, toy 273/5 (items/atlases).
C3 — validated: draws ≤350 (record 41, book 42, toy 51), determinism
0 fails, leak-free (0 geo/tex delta ×20; buy-soak 12 visits/5 bought,
leakGeo0/tex0), reload resets wallet, flag-off + ?noassets untouched,
qa.sh --strict GREEN 5/5. Shot: docs/shots/laneC/buyloop_r8.jpg.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The carried brief (slipped R7) lands complete — all four items, qa.sh --strict GREEN.
1. Selfcheck PARITY: extracted structuralSuite(plan,label) and run the FULL invariant suite
on every (source, town) — synthetic + both OSM towns get identical coverage (storeys,
one-shop-per-lot, facing, within/cross-block overlap, corridor coverage, determinism,
exactly-one-openLate). ALL GREEN 1727/1727 (was 1362; OSM went subset → full ×2 towns).
2. SECOND TOWN: Katoomba (Blue Mountains, 19 shops, op-shop/book heavy) — regional contrast
to inner-Melbourne. osm_fixture.js → OSM_TOWNS{melbourne,katoomba}+DEFAULT_TOWN;
generatePlanFor(seed, source, {town}); ?plansrc=osm&town=katoomba. Golden 0x0f652510.
Melbourne 0x34cfdec0 + synthetic 0x3fa36874 UNCHANGED (town key kept off the plan).
3. "Add a town" recipe in LANE_A_NOTES (R8): extract → OSM_TOWNS → pin golden → tell F.
4. Normalization + LOG: importer bends real data to the contracts (unknown kind→opshop,
hours-clamp so only one shop is openLate, video-preferred landmark w/ fallback) and logs
it via {report}; selfcheck prints per-town (e.g. melbourne hoursClamped:5, openLate:video).
For F: three goldens to pin, gate keyed (seed, plansrc, town); selector gained {town} — see
LANE_A_NOTES R8. Screenshots docs/shots/laneA/osm-{melbourne,katoomba}.png.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New js/world/weather.js + a one-line lighting.setSky(). Seeded per citySeed (weatherFor):
~57/24/19% clear/overcast/rain over 400 seeds, deterministic. Rain = ONE draw (camera-following
Points + procedural streak/fall/wind shader) + wet ground (roughness/darken on shared ground mats,
restored on dispose) + matching rainy sky dome via lighting.setSky. Fully procedural (0 depot fetch).
- Exposes window.PROCITY.weather = { state:'clear'|'overcast'|'rain', intensity:0..1 } for Lane D's
weather-reactive crowds (decision #1). Documented the contract + the exact 3-line shell wiring for F
in LANE_B_NOTES (I don't touch index.html). ?weather=1 seeded (default town rolls clear);
?weather=rain|overcast|clear forces a state for demos/smokes.
- Worst view 141 draws day / 126 night (<=300); composes with ?winmap + night (rain outside a lit
interior-mapped window = the money shot, screenshotted); deterministic; 0 console errors; flag-off
byte-identical (shell never constructs it). Deferred cut line: vertex-shader wind sway on
trees/awnings (touches other lanes' materials) — noted, not done. Tram stretch not attempted.
qa.sh --strict GREEN (5/5).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
qa.sh: flags harness warn_gate → run_gate — flags-off regression + the 4 existing flags +
all-on combo are now STRICT (a harness FAIL fails qa). New flags stay warn-level in-harness.
?stock=real wiring (F owns the shell + interior bridge): shell preloads the record stock pack
(preloadStockPack('record'), fail-soft, off under ?noassets) + passes stockReal to the bridge.
interior_mode builds makeStockAdapter(getStockPack(shop.type)) and passes it to BOTH buildInterior
(static crate sleeves) and dig.open (riffle) — null pack → parody canvas (Lane C fail-soft).
Harness: added stock=real smoke (assert non-parody = batched userData.isStock atlas meshes > 0;
verified 7 real sleeves in the record shop) + weather=1 smoke (auto-skips until Lane B lands),
both warn-level (decision #4). qa.sh --strict GREEN (5 gates); crates now show real covers.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The first deliberate flags-off baseline change. On D's GO (43edabf: E1 ped sub-mesh merge
447188a validated → ~1 draw/near-rig → full density 241 worst-street draws <=300, leak-free).
index.html: chunk-streamed roster default-on (chunkStream:{radius:2}); ?roster=v1 restores the
old fixed roster (escape hatch); ?roster=stream still accepted (now a no-op). Hours-aware
setNightLivelyChunks keeps the open-late block lively at night.
Flags-off regression baseline RE-PINNED (same commit, prime-law amendment):
OLD (roster-off, v1.1): 163 draws, 19000 tris
NEW (roster default-on, R7): 164 draws, ~66000 tris
Draws barely move — E1's merge makes each near-rig ~1 draw (exactly what gated the flip). The
tris jump is the streamed crowd (D worst-view 65,899); it is frame-variable, so the drift WARN
is now DRAWS-ONLY, tris guarded by the STREET_TRIS_MAX hard cap. PLAN GOLDENS UNCHANGED
(synthetic 0x3fa36874, osm 0x34cfdec0).
Also fixed the harness roster probe (read a nonexistent prop c.streaming → was blind to
streamMode) + flipped the assertion to REQUIRE streamMode on default boot + added a ?roster=v1
escape-hatch check. flags_check GREEN; qa.sh --strict green; 0 console errors both modes.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The stockAdapter seam built in v1 finally eats. Lane E's stock pack (parody-
transformed Discogs metadata + cover atlas) feeds record-shop bins + the dig
riffle behind ?stock=real. Economy stays a stub (buying = toast).
- Contract handshake (task 1, unblocked E): confirmed E's index schema as-is
in LANE_C_NOTES (items[{id,title,artist,price,price_band,atlas,uv}]; uv
origin top-left, C flips V for WebGL; fail-soft; per-type packs).
- stockpack.js (new): preloadStockPack/getStockPack (cached) + makeStockAdapter.
Atlas texture + material are shared/cached (like depot GLBs); each sleeve is
a cover PLANE sampling the atlas at its baked UV, so all real sleeves in a
room batch to ONE draw (record: 59 draws).
- stock.js binFan: real cover planes (else parody canvas). dig.js: riffles real
covers, pull card shows title/artist/$price/band. interiors.js: opts.stock=
'real' wires the adapter (warn-once + parody fallback if pack absent). Test
page reads ?stock=real, preloads record/book/toy, passes the adapter to the dig.
Validated (E's staged 24-sample; real covers drop in unchanged when E gets the
DB): determinism 0/30, leak-free rooms + real dig (0/0 over 25 cycles), draws
<=139 all rooms (record 59), fail-soft on no-pack types, ?dig=1&stock=real
leak-free. Flag-off + ?noassets untouched (procedural path byte-identical).
qa.sh --strict GREEN. F smoke hook + E contract in LANE_C_NOTES. Shot
docs/shots/laneC/stock_real_r7.jpg.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Published R5, unconsumed until now:
- glb.js KIND_TO_GLB: glassCase->glass_case, magazineRack->magazine_rack,
and a NEW crateStack primitive (fittings.js — a stack of crates) ->
crate_stack. yaw 0 (validated on-screen).
- theme.js: crateStack in record + opshop (back-room dressing); glassCase
added to milkbar; magazineRack added to pawn; milkbar magazineRack moved
window->wall (dodges the door-counter congestion).
- layout.js: fixed placeNearCounter — it picked one random side and gave up,
so counter-zone fittings (the glass case) rarely landed. Now it tries both
sides + offsets, then falls back to a free aisle spot, so a min>=1 fitting
reliably lands.
Validated (GLB on, baked depot): all 3 load + render planted/scaled/facing
right. Appearance glassCase 12/12 (toy/pawn/milkbar), magazineRack 12/12
(milkbar/pawn), crateStack ~11/12 (record/opshop). drawSweep still <=350
with them on (worst 177 GLB-on / 157 off). Determinism 0 fails, leak-free,
dig (?dig=1) still riffles batched bins (C3 re-verify). Contact sheet
docs/shots/laneC/glb_orphans_r6.jpg. qa.sh --strict GREEN.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Consumes the two orphaned street GLBs via the use-if-ready furniture path (fail-soft; ?noassets →
primitives, 0 network).
- Bins: genuinely sparse (~1/block, s+=78m, seeded side), decoupled from benches — country town not
a mall. GLB on its 0.57x0.68 footprint.
- Bus shelters ON THE STREET GRAPH (future tram/bus stops): deterministic rule = midpoint of every
MAIN edge with hashEdge(id)%3===0, road-side footpath by the kerb, long axis along the edge, open
front to road -> 1-3/town (2 in seed 20261990). Solid oriented-rect collider (furniture now returns
colliders; chunks.js merges them into getColliders) -> walk-into-it is blocked. Exported
busShelterStops(plan) -> [{edgeId,x,z,yaw}] as the stop-list contract for the vehicle loop.
- Perf: worst view WITH rig-fleet citizens 94k tris / 148 draws (<=200k / <=300, big margin);
deterministic; all-flags-on (winmap+dig+roster) composes clean, 0 errors. street_bin GLB is 3k tris
-> flagged to Lane E as a decimation candidate (bounded by the sparse cadence for now).
qa.sh --strict GREEN.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The carried-over OSM plan source (no-showed R4–R5) lands complete, invariants green — a
second CityPlan producer behind the same contract, so B–F consume it unchanged.
- osm_fixture.js: 95 REAL inner-Melbourne secondhand/charity/music/book/toy/antique shops,
extracted deterministically from thriftgod city_source.json (Overpass cache), imported as a
module ⇒ ZERO network.
- plan_osm.js: generatePlanOSM(seed) — project lat/lon→metres, bin shops into latitude-band
avenues off a spine (real row + real E–W order preserved), march into a valid non-overlapping
CityPlan, centre on origin, report true 362×486 bbox, real shop names, one openLate video.
- index.js: generatePlanFor(seed, source) selector — default 'synthetic' BYTE-IDENTICAL
(golden 0x3fa36874 untouched); 'osm' → the importer. generatePlanOSM exported for F.
- selfcheck.js: parameterized by source — OSM runs the full structural invariant suite
(frontEdge/facing/no-overlap/chunk-coverage/finite/JSON/determinism/openLate) + its own
golden 0x34cfdec0; synthetic-only brief checks skipped. ALL GREEN 1362/1362.
- map.html: ?plansrc=osm renders it (reference wiring for F). Screenshot in docs/shots/laneA/.
- CITY_SPEC: plan-source note; LANE_A_NOTES R6: seam + goldens + shape-diffs for F's gate (F2).
Honest cut: street geometry regularized (real row+order drive it, exact metres do not);
name-parody deferred. qa.sh --strict GREEN.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
C shipped dig.js + the flag but left the in-game input/mode hook to F (LANE_C_NOTES C2).
interior_mode.js (F-owned): aim at a userData.kind==='bin' fitting + press E -> createDig +
dig.open (per-bin seed binSeed(shop.seed,pos)); the dig owns the frame (renders dig.scene,
suppresses interior walk); WALK OUT/Esc closes back to the room; closed on interior exit.
Shell (index.html): ?dig=1 -> dig: opt; procity:digOpen releases pointer-lock (riffle is
cursor-driven); the unlock->leaveShop guard reads interiorMode.digActive so opening a bin
isn't read as walking out; canvas re-locks for interior walk after close.
Fresh-chromium smoke: E opens the riffle (seeded sleeves + prices render), WALK OUT closes
to the interior, leak-free (steady-state geo/tex flat over 3 open/close cycles), 0 console
errors. Flag off => no listener, dig stays null => byte-identical to v1 (prime law).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Ported 90sDJsim/web/world/dig.js → web/js/interiors/dig.js. Walk to a record
bin, press E, riffle its sleeves (scroll/drag), click the front sleeve to pull
it, BUY IT. Faithful to the source: hinge-forward riffle, pull-and-inspect
panel, procedural fwip/thunk audio.
- Riffle items = seeded procedural sleeves per bin via the stockAdapter(shop,
'sleeve') seam (parody 90s-AU records now; real GODVERSE records feed the
same seam in round 6+). Same bin ⇒ same order.
- Buy is STUBBED (toast + remove-from-bin; getCash returns a constant) —
economy is round 6+ per integrator decision #1.
- Own scene + camera; leak-free: per-open GPU resources (sleeve textures/
materials, crate parts) freed on every close(), persistent set freed on
dispose(). Dig soak (20× enter/leave) → leakGeo 0 / leakTex 0.
- Flag-gated (?dig=1): buildInterior is UNTOUCHED, so ?dig=0 boot is
byte-identical to v1 (prime law). The dig is wired only in the test page via
an E-key raycast; F owns the shell input/mode seam — hook documented in
docs/LANES/LANE_C_NOTES.md.
Verified: riffle + pull + buy-stub work; dig soak leak-free; placement
determinism unaffected; close returns to the interior. qa.sh --strict GREEN.
Shot: docs/shots/laneC/dig_riffle_r5.jpg.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Implements the R3 design note (was design-only). Default-off; the v1 whole-graph path is
byte-identical when the flag is off (shell citizen-0 signature matches the R4 golden exactly).
- chunk-local identity keyed (seed, chunkKey, i) — same seed + chunk -> same residents,
independent of town size / visit order; home edge picked from that chunk's edges.
- windowed spawn/despawn: hook-driven via Lane B's onChunkBuilt/onChunkDisposed (sim auto-
detects and stops polling) + a camera-poll fallback that needs zero Lane B changes.
- global near-cap + mixer stagger + rig pool + impostor layer unchanged, run across live chunks.
- owner-chunk despawn (no cross-boundary re-keying); per-chunk time-of-day thinning (hours-aware).
Proven on a 9x9 fixture grid: constant far-field density (~58-129 within 70m at spawn AND 150m
out, vs v1's uniform ~12-15); deterministic (507 chunk-keyed identities re-derive from seed);
leak-free (2 full grid walks / ~380 chunk cycles -> 0 geom/tex delta, pool capped 30); budget
unchanged (near-cap 24 + 1 impostor draw, 2.31ms/frame). qa.sh --strict GREEN.
Only web/js/citizens/sim.js + web/citizens_test.html changed. Flag + shell call-site documented
in LANE_D_NOTES for Lane F to wire web/index.html.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Vuntra window trick — a passing pedestrian sees a plausible room through the shop glass with real
parallax, no geometry. One shared ShaderMaterial does single-cube interior mapping on the existing
window mesh (back wall + seeded shelves + side walls + floor/ceiling); per-shop variation via
aTangent/aTint/aSeed attributes, still 1 window draw/chunk. §3.5 held via the shared uHour/uNight +
aHours: night → open room glows warm, closed room dark.
v2 prime law: WINMAP is read from the URL in buildings.js (self-contained — no index.html seam; F
just needs a flags-table row). Flag-OFF byte-identical to v1 (windowQuad → v1 hoursQuad, v1
windowMaterial). Proven: flag-off windows MeshStandardMaterial w/ no winmap attrs; flag-on interior
ShaderMaterial; worst view 122-168 draws (<=300); 100% procedural (0 fetch delta vs ?noassets
baseline); 0 console errors; GOLDEN.hash untouched. Screenshots in docs/shots/laneB/winmap_*.
qa.sh --strict GREEN (4/4).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Round-5 §Lane C task 1 (the v1.1 gate). Map Lane E's R4-late re-exports in
glb.js:
- record_crate FIXED (E stripped sourceless AO/roughness slots) → crate/
recordBin upgrade to plastic crates with the procedural sleeves inside;
removed the stale "fails to load" comment.
- counter → 'counter_till' (the real ~1.6m timber counter WITH a modelled
till; NOT the 4m balcao). Footprint-fit + floor-plant + keeper-stand pose
unchanged.
- Double-till resolved (documented): counter_till carries its own till, so
the separate cash_register attach is dropped whenever the counter GLB is
live (falls back to cash_register on the primitive counter otherwise);
milkshake_mixer stays on the milk-bar bench, lifted +0.10m onto the taller
GLB counter.
Validated (GLB on, via depot): sweep+soak {throws 0, path 0, det 0, worst
10.4ms}, 0 rooms with a visible primitive till, leak-free vs a saturated
baseline (the transient 2geo/4tex first-sight is record_crate + counter_till
cache-fill). Primitive path byte-unchanged (glb.js-only edit). qa.sh --strict
GREEN. Contact sheet docs/shots/laneC/glb_v1_1_r5.jpg. → F can tag v1.1
(LANE_C_NOTES.md).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
