jpmschweitzer/settled-reach
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(simulation): region climate Atlas layer + quarter-footprint design (T-1113/T-1112) #179

Closed
jpmschweitzer wants to merge 0 commits from atlas-layer-wiring into main
pull from: atlas-layer-wiring
merge into: :main
Conversation 3 Commits - Files Changed -
jpmschweitzer commented 2026-07-16 19:08:26 +02:00
Owner
Copy Link
Copy Source

Summary

The Atlas wiring pair. T-1113 wires the region climate stack (T-1078, implemented-but-never-called) into the cascade and serves it over the layer proxy; T-1112 records the QuarterFootprintLayer design as a D-226 amendment (Tyre data shape + Araminta encoding). Both visual overlays remain follow-up tickets by design.

T-1113 — region climate layer (server + one client line)

  • CascadeLayer::Region appended per the enum append-only Ord rule (depends only on seed/body_params/heightmap dims; documented on the variant). The region cache was already being computed and DISCARDED inside derive_all_districts — this retains it as its own layer (one-layer-one-concern; cheap double-derive accepted, documented).
  • BodyWorldState.regions: BTreeMap per the districts precedent; ~195x98 max, trivial footprint.
  • RegionGridLayer: dense row-major, all-integer wire encoding (season/weather repr(u8) discriminants; mean_temp as deci-degC i16 with i16::MIN airless sentinel; moisture_q u8). build_region_grid mirrors build_district_grid; the layer_proxy growth-ceiling doc already earmarked this ticket.
  • protocol.gd: one-line region_grid passthrough; wire fixtures regenerated via make fixtures.
  • Tests: cascade Region-layer determinism (bit-level f32 checks) + populated/skip paths; encoder unit test (dims, discriminants, rounding, sentinel, clamp); ordering pin. atlas:: suite 507 green.

T-1112 — QuarterFootprintLayer design (D-226 amendment, no code)

  • Aggregates-only, keyed by city_id: QuarterWorldState carries NO spatial position (QuarterId is a content-addressed hash) and a 512m quarter is sub-pixel at planetary projection — two independent analyses (data + rendering) converged on no-outline-by-construction.
  • Five quantized fields (density_avg_pct, dominant_district_type, dominant_zoning, landmark_count, corridor_count); D-226(d) aggregate-stats ceiling restated concretely; scaled-glyph encoding anchored on the L3 settlement dot (shape=dominant type, color=density ramp on the settlement-gold family); landmark/corridor counts tooltip-only; zoom gating reuses SETTLEMENT_LABEL_MIN_ZOOM; five wiring touch points named for the follow-up.

Also

  • test(client): input-roundtrip flaky fix — the server binds ~250ms after spawn, so a random-port collision (os error 98) escaped the spawn-liveness check and failed at the push gate; deaths during the connect window now retry on a fresh port. Targeted suite 6/6.

Verification

Full gate green: GDScript parse, client suite 2948 tests, fmt/clippy/full cargo test/deny; pql decisions validate + sync clean (387 records, 0 broken refs). No systems.db change (no stamp impact).

Follow-ups (filed at merge)

Climate overlay rendering (consumes region_grid); quarter-footprint wiring (implements the amendment); Atlas screenshot golden harness (T-1037).

## Summary The Atlas wiring pair. T-1113 wires the region climate stack (T-1078, implemented-but-never-called) into the cascade and serves it over the layer proxy; T-1112 records the QuarterFootprintLayer design as a D-226 amendment (Tyre data shape + Araminta encoding). Both visual overlays remain follow-up tickets by design. ### T-1113 — region climate layer (server + one client line) - CascadeLayer::Region appended per the enum append-only Ord rule (depends only on seed/body_params/heightmap dims; documented on the variant). The region cache was already being computed and DISCARDED inside derive_all_districts — this retains it as its own layer (one-layer-one-concern; cheap double-derive accepted, documented). - BodyWorldState.regions: BTreeMap per the districts precedent; ~195x98 max, trivial footprint. - RegionGridLayer: dense row-major, all-integer wire encoding (season/weather repr(u8) discriminants; mean_temp as deci-degC i16 with i16::MIN airless sentinel; moisture_q u8). build_region_grid mirrors build_district_grid; the layer_proxy growth-ceiling doc already earmarked this ticket. - protocol.gd: one-line region_grid passthrough; wire fixtures regenerated via make fixtures. - Tests: cascade Region-layer determinism (bit-level f32 checks) + populated/skip paths; encoder unit test (dims, discriminants, rounding, sentinel, clamp); ordering pin. atlas:: suite 507 green. ### T-1112 — QuarterFootprintLayer design (D-226 amendment, no code) - Aggregates-only, keyed by city_id: QuarterWorldState carries NO spatial position (QuarterId is a content-addressed hash) and a 512m quarter is sub-pixel at planetary projection — two independent analyses (data + rendering) converged on no-outline-by-construction. - Five quantized fields (density_avg_pct, dominant_district_type, dominant_zoning, landmark_count, corridor_count); D-226(d) aggregate-stats ceiling restated concretely; scaled-glyph encoding anchored on the L3 settlement dot (shape=dominant type, color=density ramp on the settlement-gold family); landmark/corridor counts tooltip-only; zoom gating reuses SETTLEMENT_LABEL_MIN_ZOOM; five wiring touch points named for the follow-up. ### Also - test(client): input-roundtrip flaky fix — the server binds ~250ms after spawn, so a random-port collision (os error 98) escaped the spawn-liveness check and failed at the push gate; deaths during the connect window now retry on a fresh port. Targeted suite 6/6. ### Verification Full gate green: GDScript parse, client suite 2948 tests, fmt/clippy/full cargo test/deny; pql decisions validate + sync clean (387 records, 0 broken refs). No systems.db change (no stamp impact). ### Follow-ups (filed at merge) Climate overlay rendering (consumes region_grid); quarter-footprint wiring (implements the amendment); Atlas screenshot golden harness (T-1037).
jpmschweitzer added 4 commits 2026-07-16 19:08:26 +02:00
chore(meta): activate T-1112/T-1113 atlas-layer batch — Si refinement appended (both READY) c2f236e26e
feat(simulation): wire region climate layer into the cascade + proxy (T-1113, D-243) 4f52abd532 
CascadeLayer::Region appended per the enum's append-only Ord rule (depends only on seed/body_params/heightmap dims — documented); BodyWorldState.regions BTreeMap per the districts precedent; RegionGridLayer dense row-major all-integer encoding (season/weather repr(u8) discriminants, mean_temp deci-degC i16 with i16::MIN airless sentinel, moisture_q u8); build_region_grid mirrors build_district_grid; protocol.gd region_grid passthrough (visual overlay deliberately out of scope); wire fixtures regenerated via make fixtures. atlas:: suite 507 green.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
docs(meta): D-226 amendment — QuarterFootprintLayer design (T-1112; Tyre data shape + Araminta encoding) 974897b889 
Aggregates-only layer keyed by city_id (QuarterWorldState carries no spatial position — QuarterId is a content-addressed hash; a 512m quarter is sub-pixel at planetary projection, so no outline geometry by construction). Five quantized fields (density_avg_pct, dominant_district_type, dominant_zoning, landmark_count, corridor_count); scaled-glyph encoding anchored on the L3 settlement dot (shape=dominant type, color=density ramp); landmark/corridor counts tooltip-only per the D-226(d) aggregate-stats ceiling; five wiring touch points named for the follow-up ticket.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
test(client): retry input-roundtrip server spawn on fresh port when it dies pre-connect bba14cb9ba 
The server binds AFTER ~250ms of plugin/sim init, so a random-port collision (os error 98) escapes the 0.15s spawn-liveness check and surfaced as a hard test failure at the push gate. Death during the connect window now retries on a fresh port within MAX_PORT_ATTEMPTS. Verified: targeted suite 6/6.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer commented 2026-07-16 19:16:19 +02:00
Author
Owner
Copy Link
Copy Source

Review: atlas-layer-wiring → main (type: code + design amendment)

Reviewers: Hoshe (code quality) + Troblum holding the architecture seat — Tyre authored this PR's D-226 amendment (T-1112), so the independent second opinion goes to the external sparring partner per his remit. Both read from the branch worktree at tip bba14cb9b. Gate passed pre-push (GDScript parse, 2948 client tests, fmt/clippy/full cargo test/deny) and was not duplicated.

Hoshe (Code Quality): REQUEST_CHANGES

Wiring verified sound: covering-grid math hand-traced (div_euclid, HEIGHTMAP_CELLS_PER_DISTRICT, DISTRICTS_PER_REGION=100), deci-°C i16 quantization safe against class-band extremes with the sentinel correctly reserved via clamp(i16::MIN+1, ...), fixtures/protocol.gd passthrough honest, flaky-test fix clean. Two findings:

# File Issue
H1 server/src/atlas/layer_proxy.rs:33-37 Stale doc: AtlasLayerRequest.up_to still documents run_work_item as running "through CascadeLayer::RoadGraph (the terminal layer, T-1038)" — this PR changed the terminal to Region and updated the growth-ceiling comment below but missed this one.
H2 server/src/atlas/layer_proxy.rs:1108-1184 Coverage gap: the REAL dispatch path (gen_queue up_to=Region → BodyWorldState.regions → handle_atlas_request → region_grid) is never exercised. body_params_reader_wired_produces_populated_regions asserts only districts despite its name; gen_queue's own tests always pass body_params: None. Cheap fix in existing plumbing: assert regions non-empty + cache-insert + second handle_atlas_request asserting region_grid.is_some(); sibling no-params test asserts regions empty.

Troblum (Architecture, independent): REQUEST_CHANGES

The T-1112 amendment's core claims verified independently: QuarterWorldState/QuarterSkeleton carry no spatial position (every field read), the city_id-recompute join is a real pure function, the sub-pixel scale math checks out, and every cited encoding convention exists as described. T-1113's tail-append ordering, double-derive tradeoff, and f32-internal/i16-wire split all ruled sound (D-227/D-010-consistent). Three findings:

# File Issue
T1 governance/decisions/architecture.md (T-1112 amendment §1, ~1581) The dominant_zoning tie-break ("lowest declaration-order variant") is not computable: ZoningType derives no Ord/PartialOrd (unlike DistrictType, which has them per its T-994 precedent). The §4 touch-point list must name the derive addition (or pin an Ord-free rule) — otherwise the wiring implementer diverges from the record silently.
T2 server/src/atlas/layer_proxy.rs:33-37 Same stale terminal-layer doc as H1 (found independently).
T3 server/src/atlas/believability.rs:506-535 cascade_for_body still hardcodes CascadeLayer::RoadGraph while production dispatch now runs to Region — silent harness/production depth divergence. Harmless today (verified: believability never reads state.regions), but D-245's climate-appropriateness condition is a natural future consumer that would silently see empty regions. Bump to Region (mirror production) or record the lag deliberately.

Verdict: CHANGES REQUESTED

4 unique findings (H1≡T2). All small and precisely located; fix round dispatched with the ruling on T3 = bump to mirror production.

## Review: atlas-layer-wiring → main (type: code + design amendment) Reviewers: Hoshe (code quality) + **Troblum** holding the architecture seat — Tyre authored this PR's D-226 amendment (T-1112), so the independent second opinion goes to the external sparring partner per his remit. Both read from the branch worktree at tip bba14cb9b. Gate passed pre-push (GDScript parse, 2948 client tests, fmt/clippy/full cargo test/deny) and was not duplicated. ### Hoshe (Code Quality): REQUEST_CHANGES Wiring verified sound: covering-grid math hand-traced (div_euclid, HEIGHTMAP_CELLS_PER_DISTRICT, DISTRICTS_PER_REGION=100), deci-°C i16 quantization safe against class-band extremes with the sentinel correctly reserved via clamp(i16::MIN+1, ...), fixtures/protocol.gd passthrough honest, flaky-test fix clean. Two findings: | # | File | Issue | |---|------|-------| | H1 | server/src/atlas/layer_proxy.rs:33-37 | Stale doc: AtlasLayerRequest.up_to still documents run_work_item as running "through CascadeLayer::RoadGraph (the terminal layer, T-1038)" — this PR changed the terminal to Region and updated the growth-ceiling comment below but missed this one. | | H2 | server/src/atlas/layer_proxy.rs:1108-1184 | Coverage gap: the REAL dispatch path (gen_queue up_to=Region → BodyWorldState.regions → handle_atlas_request → region_grid) is never exercised. body_params_reader_wired_produces_populated_regions asserts only districts despite its name; gen_queue's own tests always pass body_params: None. Cheap fix in existing plumbing: assert regions non-empty + cache-insert + second handle_atlas_request asserting region_grid.is_some(); sibling no-params test asserts regions empty. | ### Troblum (Architecture, independent): REQUEST_CHANGES The T-1112 amendment's core claims verified independently: QuarterWorldState/QuarterSkeleton carry no spatial position (every field read), the city_id-recompute join is a real pure function, the sub-pixel scale math checks out, and every cited encoding convention exists as described. T-1113's tail-append ordering, double-derive tradeoff, and f32-internal/i16-wire split all ruled sound (D-227/D-010-consistent). Three findings: | # | File | Issue | |---|------|-------| | T1 | governance/decisions/architecture.md (T-1112 amendment §1, ~1581) | The dominant_zoning tie-break ("lowest declaration-order variant") is not computable: ZoningType derives no Ord/PartialOrd (unlike DistrictType, which has them per its T-994 precedent). The §4 touch-point list must name the derive addition (or pin an Ord-free rule) — otherwise the wiring implementer diverges from the record silently. | | T2 | server/src/atlas/layer_proxy.rs:33-37 | Same stale terminal-layer doc as H1 (found independently). | | T3 | server/src/atlas/believability.rs:506-535 | cascade_for_body still hardcodes CascadeLayer::RoadGraph while production dispatch now runs to Region — silent harness/production depth divergence. Harmless today (verified: believability never reads state.regions), but D-245's climate-appropriateness condition is a natural future consumer that would silently see empty regions. Bump to Region (mirror production) or record the lag deliberately. | ### Verdict: CHANGES REQUESTED 4 unique findings (H1≡T2). All small and precisely located; fix round dispatched with the ruling on T3 = bump to mirror production.
jpmschweitzer added 1 commit 2026-07-16 19:22:24 +02:00
fix(simulation): PR #179 review round — F1-F4 9d92d8090f 
F1: T-1112 amendment touch-point list gains the ZoningType PartialOrd/Ord derive (mirroring DistrictType's T-994 precedent) so the dominant_zoning declaration-order tie-break is computable. F2: stale terminal-layer doc on AtlasLayerRequest.up_to corrected to Region. F3: believability cascade_for_body bumped to Region — harness mirrors production depth, invariant documented. F4: end-to-end dispatch coverage — populated-regions test asserts regions + second handle_atlas_request returns region_grid; no-params sibling asserts regions empty. Suites: layer_proxy 10, believability 4, gen_queue 8, cascade 10.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
jpmschweitzer commented 2026-07-16 19:22:45 +02:00
Author
Owner
Copy Link
Copy Source

Review response — all 4 findings fixed (commit 9d92d8090)

No pushbacks.

# Finding Resolution
T1 dominant_zoning tie-break not computable (ZoningType lacks Ord) Fixed — the T-1112 amendment's touch-point list (now six) names the enabling change: ZoningType gains PartialOrd/Ord mirroring DistrictType's T-994 precedent. Tie-break rule untouched; design-only framing preserved (derive is a named follow-up touch point, not built here). pql decisions validate ok.
H1/T2 Stale terminal-layer doc on AtlasLayerRequest.up_to Fixed — now documents CascadeLayer::Region as terminal (T-1113).
T3 Believability harness depth diverged from production Fixed per ruling — cascade_for_body bumped to Region; the mirror-production invariant is documented on the function (naming D-245 climate-appropriateness as the future state.regions consumer). Believability suite 4/4.
H2 No end-to-end dispatch coverage for regions Fixed — body_params_reader_wired_produces_populated_regions now asserts regions non-empty AND round-trips a second handle_atlas_request from the cache asserting Ready + region_grid.is_some(); the no-params sibling asserts regions stay empty. layer_proxy 10/10.

Gate re-run green on push (parse, full client suite, fmt/clippy/full cargo test/deny). Reviewers asked to confirm.

## Review response — all 4 findings fixed (commit 9d92d8090) No pushbacks. | # | Finding | Resolution | |---|---------|------------| | T1 | dominant_zoning tie-break not computable (ZoningType lacks Ord) | Fixed — the T-1112 amendment's touch-point list (now six) names the enabling change: ZoningType gains PartialOrd/Ord mirroring DistrictType's T-994 precedent. Tie-break rule untouched; design-only framing preserved (derive is a named follow-up touch point, not built here). pql decisions validate ok. | | H1/T2 | Stale terminal-layer doc on AtlasLayerRequest.up_to | Fixed — now documents CascadeLayer::Region as terminal (T-1113). | | T3 | Believability harness depth diverged from production | Fixed per ruling — cascade_for_body bumped to Region; the mirror-production invariant is documented on the function (naming D-245 climate-appropriateness as the future state.regions consumer). Believability suite 4/4. | | H2 | No end-to-end dispatch coverage for regions | Fixed — body_params_reader_wired_produces_populated_regions now asserts regions non-empty AND round-trips a second handle_atlas_request from the cache asserting Ready + region_grid.is_some(); the no-params sibling asserts regions stay empty. layer_proxy 10/10. | Gate re-run green on push (parse, full client suite, fmt/clippy/full cargo test/deny). Reviewers asked to confirm.
jpmschweitzer commented 2026-07-16 19:24:35 +02:00
Author
Owner
Copy Link
Copy Source

Re-review verdict: APPROVED — Hoshe: both findings verified by execution (layer_proxy 10/10, cascade 10/10; full dispatch-to-region_grid loop closed). Troblum: all three resolved — the ZoningType derive is correctly record-only (design-only discipline held), stale doc fixed, believability harness now mirrors production depth with a forward-maintenance note. No new concerns; untouched files byte-identical to round one. Merging.

## Re-review verdict: APPROVED — Hoshe: both findings verified by execution (layer_proxy 10/10, cascade 10/10; full dispatch-to-region_grid loop closed). Troblum: all three resolved — the ZoningType derive is correctly record-only (design-only discipline held), stale doc fixed, believability harness now mirrors production depth with a forward-maintenance note. No new concerns; untouched files byte-identical to round one. Merging.
jpmschweitzer closed this pull request 2026-07-16 19:30:12 +02:00

Pull request closed

This pull request cannot be reopened because the branch was deleted.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
No items
No labels
Milestone
No items
No Milestone
Projects
Clear projects
No projects
Assignees
Clear assignees
jpmschweitzer (Jeroen Schweitzer)
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: jpmschweitzer/settled-reach#179