Implement the D-230 on-demand derive phase (data-production half). New
atlas/shell.rs derives a sparse building shell for one 64 m chunk from the
plan-phase BuildingPropertyTags: rectangle-containment (footprint ∩ chunk) ×
z-range (per-floor voxel bands from FloorExtent), in a quarter-ground z frame
(D-110). Walls on the footprint perimeter, FloorSlab on interior floor bases,
Roof above the top floor; interior air is Void and never stored.
- gen_queue.rs: GenWorkItem::FillChunk carries pre-resolved block_tags +
block_pos + sub_chunk (run_work_item stays cache-free, mirroring
GenerateSkeleton); GenCompletion::ChunkFilled carries Box<FilledChunk>; the
arm calls shell::fill_chunk; pure build_fill_chunk_item(&QuarterWorldState,..)
added.
- plugin.rs: ChunkFilled handler accepts the shell (trace-only — the Phase-5
rendering consumer and the on-demand streaming dispatch trigger are deferred;
do not build on legacy chunk_streaming.rs before Phase 5, gated by T-962).
Tests: determinism, sparsity-vs-dense-volume, wall/floor/roof correctness,
basement z-origin, sub-chunk clipping, full submit→drain→ChunkFilled round-trip.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Separated the asset pipeline from the cascade work. T-1049/T-1050 assume a sprite-centric
catalog; direction is 3D objects (Trellis .glb) in-world with textures as the only
flattened content. Held for re-scope. Cascade track (T-1044/T-1047) continues.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two parallel tracks with refinement appends: T-1044+T-1047 compose on the run_layer1
pass (TerrainAnalysis cache + per-district D8 basin_direction); T-1049+T-1050 are the
asset-pipeline track (per-class story breakdown + catalog/manifest/conventions + glb-gen
path). Staleness corrected (T-1047: basin_direction threads through DistrictProfile, not
the post-T-1077 RegionProfile).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
/whats-next batch activation for the D-243 region-tier finish.
Refinement briefs (Si pass) appended to T-1042 and T-1078 reconciling
their 2026-06-12 scope against the post-T-1077 tree. T-1046 (Region→
Chunk→Voxel wiring) flipped to done — already committed in cfd5e18d0.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Server half of wiring the derived tier into production. The D-225 proxy
(AtlasLayerResponse) carried only Layer1Output, so the Atlas could not show the
DistrictProfile tier that the cascade derives. Adds DistrictGridLayer (cols/rows
+ row-major MorphologyZone discriminants + elev_q for relief shading) and
build_district_grid(), populated on a cache hit. Because the response is
rmp_serde msgpack end-to-end, the new field reaches the client automatically.
This is the planetary-scale coarse grid (the Atlas map view); the on-demand 2km
derive_district (T-1077) is for in-world Phase 5, not the map. MorphologyZone
gains Copy (fieldless repr(u8) enum; additive).
Next: client decode (protocol.gd) + the D-226 generation overlay in the Atlas
viewer.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
First phase of the D-243 re-scale. Adds server/src/atlas/scale.rs as the
single source of truth for the metric containment ladder (voxel 1m -> chunk
64m -> block 128m -> quarter 512m -> district 2048m -> region 204.8km), the
fixed integer addressing (ChunkPos -> DistrictPos -> RegionPos), and the
elastic seam regions_per_equator(R). Compile-time const asserts enforce the
nesting so the rungs cannot silently drift again — the structural fix for the
Q-110 failure mode (three files, three different region sizes).
Wires body_radius_km (D-204) through BodyParams + the reader: the single
body-specific input to the elastic seam.
No behavior change — foundation only. The rename/re-scale of the carrier
(RegionProfile -> DistrictProfile) and on-demand district derivation follow.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Resolves Q-110 (region/chunk physical-scale anchoring), which had three
contradictory metres-per-scale assumptions in the cascade and no record
fixing metres-per-heightmap-pixel.
D-243 establishes a nested absolute-metre containment ladder with a single
elastic seam at planetary scale:
voxel 1m -> chunk 64m -> block 128m -> quarter 512m -> district 2km
-> region ~205km (100x100 districts) -> (elastic) -> planet
A body holds round(2*pi*R/204.8km) regions (the only per-body-floating
quantity); everything below is fixed integer math, so ChunkPos->RegionPos
is body-independent. The region is the largest hard block: the climate/
weather/season lockdown scale (the Q-105 cheap-dynamism source) and the
sane planetary grid. Sub-heightmap detail is invented deterministically
(interpolation + domain warp + detail-scatter, never stored, D-227); climate
is edge-fuzzed so the grid never shows ('climate does not change on a line').
Vocabulary locked and recorded in CLAUDE.md: 'region' = the 205km metric
cell only; the old 1km RegionProfile is dropped onto the 2km district; D-201
tier-4 'Region' renamed Province (the watershed/political overlay), amended
here. Refines D-239 §2 (district temperature now modulates a region baseline).
Implementation tickets: T-1077 (re-scale the code to the ladder + elastic-seam
resampling; blocks T-1046) and T-1078 (region climate stack + edge fuzz).
T-1046 is un-gated on Q-110 and now blocked only by T-1077.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Reviewing the T-1038 road-graph hub model surfaced that corp HQs are
cross-referenced into atlas_city_names as reserved rows — one per corp —
producing duplicate co-named "cities" (e.g. 10 'Groombridge' rows on
GJ380c) that the spatial layers place as separate settlements.
D-242 resolves the data model: corp HQs are not cities. The HQ relationship
lives with the corp; placement is a baked, specialization-keyed preference
(CityTenant vs Standalone, since corp_type is degenerate); population and
settlement_class are install-baked with only position seed-derived. This
supersedes the D-223 corp-HQ-cross-ref and deferred-pop/class notes.
Companion tickets under T-750: T-1074 (corp-HQ model / remove the cross-ref /
specialization map / standalone HQ settlements), T-1075 (bake per-settlement
population + settlement_class), T-1076 (road-graph hub refinement — scaled-cap
hubs, co-location collapse, hybrid minor attach, deferred RailHeadFacing pass;
blocked on T-1074 + T-1075). T-1038 annotated with shipped-vs-deferred.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
import_economics.py 2,620 → 309 lines — a thin orchestrator keeping the
exact CLI, single-transaction/rollback contract, and exit codes. The 16
import steps, MIGRATION_SQL, brands shell-out, validators, and stamp
write now live in tooling/economy-db/economy_import/ (db, migration,
economy, corporations, brands, bodies, atlas, specialization, traits,
validators, stamp, paths, errors). Full type hints throughout.
tooling/generator_sources.py replaces the triplicated source registry
(importer / stamp checker / pr-process watch list — the skill now derives
its list via --list). The registry stamps itself, and economy_import/
modules are globbed fail-closed, so a future module is stamped the moment
it exists — closing the silently-weakened-stamp failure mode.
Rider: connector config helpers centralized in tooling/db/common.py.
Byte-identical behavior proven: full-import table dump diff EMPTY over
107,843 lines / 37 tables (volatile timestamp fields excluded); dry-run
output parity; generated_brands.toml sha unchanged. make test-tooling
PASS; ruff clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
pre-push now runs the full gdUnit4 suite via tests/run-godot when client/
changed — blocking, ~100s, suite made green by T-973 and independently
re-verified. The tooling gate (make test-tooling on TOOLING_CHANGED, T-1066)
rides in the same hook. D-067 amended: the shipped chime is a deliberate
800ms synthesis (#327, ui.md UI-005); prose said 300-400ms. Ticket closes
for the maintenance wave in the changelog.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Production fixes surfaced by honest test triage:
- hud_groups.gd: _set_group_z crashed on freed HUD nodes — the typed loop
variable errors before the is_instance_valid guard runs; prune first
- fog_state.gd: _resize cleared _prev_visible (world-space keys survive
resizes), so pre-resize tiles never decayed VISIBLE→EXPLORED (D-059)
Test debt (T-928/929/934/935/936/937/938/939, T-864, T-973): lambda
local-capture bugs rewritten with array captures (now assert exact
emission counts), e2e suites updated to the current handshake +
StartupMessage protocol and stream-aware reads against the live binary,
fog perf test measures steady state, chime test pins the shipped 800ms
catalog asset (D-067 amended separately), monologue gdUnit4 API typo,
battery-warning tests follow the MetaScreen on_open lifecycle. 3 sprint2
proof tests revived (corner_reveal had passed from the wrong tile — NPC3
blocks (18,14); route corrected). Soft-skips converted to real do_skip
reporting. T-1068: 7 orphan .gd.uid deleted, _format_pop/_format_radius
deduped into atlas_format.gd (preload, no class_name — headless cache).
Suite: 1264 cases/20 failures → 1268/0, independently re-verified
(2536/2536, exit 0).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
FrameAccumulator state machine in framing.rs holds partial prefix/payload
bytes across non-blocking receive() calls — a frame split across TCP
segments no longer desyncs the stream (read_exact previously discarded
partially-consumed bytes on WouldBlock). receive_bridge_inputs now drains
all ready frames per tick (capped) instead of exactly one, covering input
batches + atlas requests in the same window.
Review hardening: EOF mid-frame escalates to Disconnected like clean EOF
(peer died with a truncated stream) instead of logging an Io error every
tick. Tests: frame split inside prefix / inside payload, multi-frame
drain, mid-frame-EOF disconnect. Corrupt-stream escalation tracked as
T-1072.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Rewrites the committed .pql/changelog/ from the old T-NNN-keyed format to
the record_id identity model (pql 1.10.0, D-26): tickets keyed by a ULID
record_id, the friendly T-NNN in ticket_idmap, references and hashes
recomputed. Produced by the one-off cmd/migrate-ids in the pql repo.
Other clones: install pql >= 1.10.0, then rm .pql/pql.db && pql plan
rebuild to regenerate the local DB from this changelog.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The four voice::lookup tests shared one on-disk cache directory (base_dir
temp/sr-voice-lookup-test + world_seed 99). VoiceCacheStore persists on Drop
(save_all) and reloads on first zone access (load_zone), and cache_hit stores
'Voiced line.' under the identical (zone_id=100, CacheKey) that cache_miss looks
up — so under parallel execution the hit test's Drop/save could leak into the
miss test's lookup, returning 'Voiced line.' instead of 'Base line.'.
Give each test a directory keyed by test label + process id, so neither parallel
tests nor concurrent cargo test runs collide. Verified clean across repeated runs.
Closes T-1035.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
derive_moisture_q matched hydrosphere on 'ocean'/'rivers'/'subsurface', but the
real bodies.hydrosphere vocab in systems.db is dominated by 'liquid_water' (175
bodies) — so most surface-water worlds fell to the '_ => 30' default, propagating
wrong precipitation_class / glaciation_grade / vegetation. Same root cause as
D-240 (code keyed on a vocab the data doesn't use).
Align the match to the canonical vocab, grouped by available surface moisture
(mirrors the [hydrosphere_maritime] table from T-1033): surface liquid
(liquid_water/ocean/ocean-coastal/extensive) -> 80; rivers/rivers-lakes/moderate
-> 55; ice -> 20; subsurface_liquid -> 15; subsurface/subsurface_ice -> 10;
minimal/trace -> 5; none -> 0. Add a liquid_water regression test and broaden the
clamp sweep to the full vocabulary.
Closes T-1034 — last open ticket under epic T-974 (Atlas-to-tile derivation),
which is now complete.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Record D-240: a body's orbital position and stellar assignment are placeholder
registry data with no authored physical intent, so no derivation may use them as
physical inputs. Climate derives from authored planet_class (temperature envelope)
+ atmosphere + hydrosphere + seed. Supersedes the sun-driven Stefan-Boltzmann
portion of D-239 §2, which produced +356 C 'temperate' worlds (NULL spectral_class
-> G-Sun default on an M-dwarf catalog).
File T-1033 under T-974: rewrite derive_temperature_c to the class-envelope model,
drop the orbit/star inputs, regen the T-1031 golden. Audit confirmed the blast
radius is one function (derive_temperature_c).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
D-239 three-carrier foundation merged to main. Review found a correctness
bug (per-region elevation) + hollow determinism tests, both fixed before merge.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
T-1023/1024/1026 implemented and pushed as PR #158; moved to review.
T-1032 (new, under T-974) tracks wiring the production AnalyzeBody dispatch
to read BodyParams from the DB — currently always None, so the carrier layer
only runs in tests.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PR #157 merged -> T-1022 done. Activate the D-239 three-carrier foundation
batch via /whats-next: T-1023 (RegionProfile), T-1026 (domain warp), and
T-1024 (district climate) with refinement notes. T-1024 blocked-by T-1023
(temperature fields live on the RegionProfile struct).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
T2: background agent / lead writes are confined to the session root, so an
out-of-root sibling worktree (.worktrees/<branch>) is unwritable — the
agent could design but not write, and there's no permission to grant. Fix
options for clide noted. Plus a positive: clide image show works first-try.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
