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

feat(simulation): L3→L4 dispatch threading — political_archetype / morphology_zone / road_entry_directions (T-1039, T-1043) #168

Closed
jpmschweitzer wants to merge 0 commits from l3-l4-dispatch-threading into main
pull from: l3-l4-dispatch-threading
merge into: :main
Conversation 2 Commits - Files Changed -
jpmschweitzer commented 2026-06-16 12:30:30 +02:00
Owner
Copy Link
Copy Source

Completes the GenerateSkeleton dispatch that T-1022 wired (it carried only founding_orientation). build_skeleton_work_item now also threads the remaining Layer-3 character into Layer 4.

T-1039 — political_archetype + morphology_zone + arrangement_pattern

  • political_archetype: real value from CityPlacement (was the Commission stub).
  • morphology_zone: looked up from the covering DistrictProfile via new scale::heightmap_pixel_to_district (exported HEIGHTMAP_CELLS_PER_DISTRICT; cascade.rs drops its local literal for the shared const). Was the AlluvialPlain stub.
  • arrangement_pattern: re-derived at L4 via the single pure fn attractor_matching::arrangement_pattern(&archetype, &economic_role) (both inputs already at L4) — no CityGenerationContext field, no stored derived state (DB-as-cache). A pure total fn, so re-derivation is provably identical to L3. Parity test asserts L4 re-derivation == the L3-stored value (drift tripwire).

T-1043 — road_entry_directions

  • Derived from T-1038's BodyWorldState.road_graph: octant = bearing of each incident road edge, deduped per octant, ordered by MaintenanceAuthority rank (the AdminFacing prestige-edge consumer). Feeds derive_access_points → AccessKind::QuarterEdge; the BlockJunction fallback now fires only for genuinely isolated settlements.

Acceptance

  • Corporate + Fjord placement → Ribbon topology + CompanyTown layout (not mesh + Commission).
  • Settlement with a road connection → QuarterEdge on the correct octant; isolated → BlockJunction.

Verification

  • cargo test (full): 1559 lib + 37 harness + all integration — 0 failures. 11 new tests (incl. the arrangement_pattern parity tripwire + both acceptance tests). clippy --all-targets -D warnings clean; fmt clean. Coordinate convention (row/col → DistrictPos) verified against derive_all_districts.

Notes for review

  • The .config/hooks/post-checkout commit (d2cbdc243) is already on main (cherry-picked as 34ffcda86) and will no-op at merge — not part of this review.
  • .pql/changelog is board meta (T-1039/T-1043 activation + T-960→backlog, T-1038→done).

🤖 Generated with Claude Code

Completes the GenerateSkeleton dispatch that T-1022 wired (it carried only `founding_orientation`). `build_skeleton_work_item` now also threads the remaining Layer-3 character into Layer 4. ## T-1039 — political_archetype + morphology_zone + arrangement_pattern - **political_archetype**: real value from `CityPlacement` (was the `Commission` stub). - **morphology_zone**: looked up from the covering `DistrictProfile` via new `scale::heightmap_pixel_to_district` (exported `HEIGHTMAP_CELLS_PER_DISTRICT`; `cascade.rs` drops its local literal for the shared const). Was the `AlluvialPlain` stub. - **arrangement_pattern**: **re-derived at L4** via the single pure fn `attractor_matching::arrangement_pattern(&archetype, &economic_role)` (both inputs already at L4) — no `CityGenerationContext` field, no stored derived state (DB-as-cache). A pure total fn, so re-derivation is provably identical to L3. **Parity test** asserts L4 re-derivation == the L3-stored value (drift tripwire). ## T-1043 — road_entry_directions - Derived from T-1038's `BodyWorldState.road_graph`: octant = bearing of each incident road edge, deduped per octant, ordered by `MaintenanceAuthority` rank (the AdminFacing prestige-edge consumer). Feeds `derive_access_points` → `AccessKind::QuarterEdge`; the `BlockJunction` fallback now fires only for genuinely isolated settlements. ## Acceptance - Corporate + Fjord placement → Ribbon topology + CompanyTown layout (not mesh + Commission). - Settlement with a road connection → QuarterEdge on the correct octant; isolated → BlockJunction. ## Verification - `cargo test` (full): 1559 lib + 37 harness + all integration — 0 failures. 11 new tests (incl. the arrangement_pattern parity tripwire + both acceptance tests). clippy `--all-targets -D warnings` clean; fmt clean. Coordinate convention (row/col → DistrictPos) verified against `derive_all_districts`. ## Notes for review - The `.config/hooks/post-checkout` commit (`d2cbdc243`) is **already on main** (cherry-picked as `34ffcda86`) and will no-op at merge — not part of this review. - `.pql/changelog` is board meta (T-1039/T-1043 activation + T-960→backlog, T-1038→done). 🤖 Generated with [Claude Code](https://claude.com/claude-code)
jpmschweitzer added 3 commits 2026-06-16 12:30:30 +02:00
feat(simulation): thread political_archetype/morphology_zone/road_entry_directions through L3->L4 dispatch (T-1039, T-1043) 9c8ad83071 
Completes the GenerateSkeleton dispatch T-1022 wired (which only carried
founding_orientation). build_skeleton_work_item now also threads:

T-1039:
- political_archetype — real value from CityPlacement, replacing the Commission stub.
- morphology_zone — looked up from the covering DistrictProfile via new
  scale::heightmap_pixel_to_district (exported HEIGHTMAP_CELLS_PER_DISTRICT;
  cascade.rs drops its local CELLS_PER_REGION literal for the shared const).
- arrangement_pattern — RE-DERIVED at L4 (option b) via the single pure fn
  attractor_matching::arrangement_pattern(&archetype, &economic_role); both
  inputs already at L4. No CityGenerationContext field, no stored derived state
  (DB-as-cache). Parity test asserts L4 re-derivation == L3-stored value.

T-1043:
- road_entry_directions — derived from T-1038's BodyWorldState.road_graph:
  octant = bearing of each incident road edge, deduped per octant, ordered by
  MaintenanceAuthority rank (the AdminFacing prestige-edge consumer). Feeds
  derive_access_points -> AccessKind::QuarterEdge; BlockJunction fallback now
  fires only for genuinely isolated settlements.

Acceptance covered: Corporate+Fjord -> Ribbon topology (not mesh+Commission);
settlement with a road -> QuarterEdge on the correct octant. 11 new tests.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
chore(meta): activate T-1039/T-1043 + board cleanup (T-960->backlog, T-1038->done) a3c3929c4e 
T-1039/T-1043 in_progress + refinement appends (the locked arrangement_pattern
re-derive + RoadGraph-source decisions). Board hygiene: T-960 (Atlas map viewer)
parked since 2026-06-07 -> backlog; T-1038 (Layer 2 road graph) full scope landed
on main (c0d4ac89d) and consumed by T-1043 -> done.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
fix(config): post-checkout preserves uncommitted pql mutations on branch creation d2cbdc2431 
The post-checkout hook ran 'pql plan rebuild' (drop + replay changelog) on every
branch checkout, including 'git checkout -b'. That silently dropped any pql.db
ticket mutation not yet in the committed changelog — e.g. an in_progress status set
right before cutting a feature branch (hit this session: T-1039/T-1043 reverted to
backlog at branch creation).

Git has no pre-checkout/pre-worktree hook, so capture the changelog at post-checkout
instead: when prev-HEAD == new-HEAD (a branch *creation* — content unchanged, pql.db
already correct), run 'pql plan export --stage' (the same flush the pre-commit hook
does) instead of the destructive rebuild. Real branch *switches* (prev != new) still
rebuild to match the branch's changelog.

Verified: simulating a same-HEAD checkout with an uncommitted in_progress mutation now
preserves it (previously reverted).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
jpmschweitzer commented 2026-06-16 12:38:24 +02:00
Author
Owner
Copy Link
Copy Source

PR #168 Review — l3-l4-dispatch-threading → main (code)

Reviewers: Hoshe (code quality) + Tyre (architecture), both on the branch via git show. Reviewed server/src/atlas/{plugin.rs, scale.rs, cascade.rs} (the hook commit is already on main; .pql/changelog is meta).

Verdict: CHANGES REQUESTED.

Tyre (Architecture) — APPROVE

Decision-consistent and determinism-clean. Verified: D-234a topology now gated by the real morphology_zone (fjord → Ribbon, not the AlluvialPlain mesh — the corporate_fjord_… acceptance test exercises it); political_archetype per D-214; arrangement_pattern re-derivation is sound (pure total fn over (archetype, economic_role), both from the same read-only DB row → can only diverge on a forbidden runtime write; parity test enforces it); scale.rs is the correct home for heightmap_pixel_to_district (avoids a district_profile circular-dep); RoadGraph dependency degrades gracefully to an empty graph (→ BlockJunction, same as isolated); D-010 maintained (BTreeMap, integer math, total sort order). Non-blocking: the arrangement_pattern doc comment in build_skeleton_work_item says "re-derived at L4" but the re-derivation happens inside generate_quarter_skeleton, not there — a one-line clarification would close the gap.

Hoshe (Code Quality) — REQUEST_CHANGES

# File / location Issue
H1 plugin.rs — build_skeleton_work_item, morphology fallback (~238-241) The AlluvialPlain fallback fires silently when districts.get(&district_pos) misses. This happens (a) in production when a body's district layer was never built (params-missing) — every city on that body silently gets wrong street topology — or (b) from a coordinate bug. No log either way, inconsistent with the surrounding tracing::warn! on the read_set path. Add tracing::debug! when the grid is empty (expected: layer absent) and tracing::warn! when the grid is non-empty but the key misses (indicates a pixel→district convention bug).
H2 plugin.rs — arrangement_pattern_l4_rederivation_matches_l3_stored_value (~456-537) The parity tripwire covers transit_hub only with Commission + Corporate. The transit_hub override fires before the archetype match (currently archetype-independent) — but that's exactly what the tripwire guards. Add transit_hub rows for Pioneer / Industrial / Military / Academic so an accidental archetype-conditionalization is caught.

Hoshe confirmed clean: the row/col→DistrictPos convention matches derive_all_districts, bearing_octant integer math (axis + diagonal boundaries handled, no atan2), deterministic dedup/ordering (BTreeMap + total tie-break), no panic paths, D-010 throughout.

Disposition: both H1/H2 + Tyre's doc note will be addressed (small additive changes).

## PR #168 Review — `l3-l4-dispatch-threading` → main (code) Reviewers: **Hoshe** (code quality) + **Tyre** (architecture), both on the branch via `git show`. Reviewed `server/src/atlas/{plugin.rs, scale.rs, cascade.rs}` (the hook commit is already on main; `.pql/changelog` is meta). **Verdict: CHANGES REQUESTED.** ### Tyre (Architecture) — APPROVE Decision-consistent and determinism-clean. Verified: D-234a topology now gated by the real `morphology_zone` (fjord → Ribbon, not the AlluvialPlain mesh — the `corporate_fjord_…` acceptance test exercises it); political_archetype per D-214; arrangement_pattern re-derivation is sound (pure total fn over `(archetype, economic_role)`, both from the same read-only DB row → can only diverge on a forbidden runtime write; parity test enforces it); `scale.rs` is the correct home for `heightmap_pixel_to_district` (avoids a district_profile circular-dep); RoadGraph dependency degrades gracefully to an empty graph (→ BlockJunction, same as isolated); D-010 maintained (BTreeMap, integer math, total sort order). **Non-blocking:** the `arrangement_pattern` doc comment in `build_skeleton_work_item` says "re-derived at L4" but the re-derivation happens inside `generate_quarter_skeleton`, not there — a one-line clarification would close the gap. ### Hoshe (Code Quality) — REQUEST_CHANGES | # | File / location | Issue | |---|-----------------|-------| | H1 | `plugin.rs` — `build_skeleton_work_item`, morphology fallback (~238-241) | The `AlluvialPlain` fallback fires silently when `districts.get(&district_pos)` misses. This happens (a) in production when a body's district layer was never built (params-missing) — *every* city on that body silently gets wrong street topology — or (b) from a coordinate bug. No log either way, inconsistent with the surrounding `tracing::warn!` on the read_set path. Add `tracing::debug!` when the grid is empty (expected: layer absent) and `tracing::warn!` when the grid is non-empty but the key misses (indicates a pixel→district convention bug). | | H2 | `plugin.rs` — `arrangement_pattern_l4_rederivation_matches_l3_stored_value` (~456-537) | The parity tripwire covers `transit_hub` only with Commission + Corporate. The `transit_hub` override fires before the archetype `match` (currently archetype-independent) — but that's exactly what the tripwire guards. Add `transit_hub` rows for Pioneer / Industrial / Military / Academic so an accidental archetype-conditionalization is caught. | Hoshe confirmed clean: the row/col→DistrictPos convention matches `derive_all_districts`, `bearing_octant` integer math (axis + diagonal boundaries handled, no atan2), deterministic dedup/ordering (BTreeMap + total tie-break), no panic paths, D-010 throughout. Disposition: both H1/H2 + Tyre's doc note will be addressed (small additive changes).
jpmschweitzer added 2 commits 2026-06-16 12:52:12 +02:00
fix(simulation): address PR #168 review — fallback diagnostic, parity coverage, doc (T-1039) 43aadc4491 
Hoshe H1: the morphology_zone AlluvialPlain fallback fired silently. Now logs
tracing::debug! when the district grid is empty (expected: params-missing body /
early cascade) and tracing::warn! when a populated grid misses the key (indicates a
pixel->district convention bug) — no more silently-wrong topology.

Hoshe H2: the arrangement_pattern parity tripwire only covered transit_hub with
Commission + Corporate. Added Pioneer/Industrial/Military/Academic rows so an
accidental archetype-conditionalization of the cross-archetype override is caught.

Tyre (non-blocking): clarified the build_skeleton_work_item doc — the
arrangement_pattern re-derivation call lives in generate_quarter_skeleton (the
consumer), not in this function.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
chore(config): align branch post-checkout with main's do-nothing version 2d85f8ca3c 
Reconciles the branch's earlier export --stage hook (d2cbdc243) with main's
superseding do-nothing fix (b24a3596b) so the two merge cleanly instead of
conflicting on this file. (The hook fix is already live on main; this just keeps
the branch copy identical.)

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
jpmschweitzer commented 2026-06-16 12:55:37 +02:00
Author
Owner
Copy Link
Copy Source

Review findings addressed — 2d85f8ca3

All findings resolved. Gate re-run green: cargo test 1559+ lib + integration (0 failures), clippy --all-targets -D warnings clean, fmt clean.

# Finding Resolution
H1 Silent AlluvialPlain morphology fallback Now tracing::debug! when the district grid is empty (expected: params-missing body / early cascade) and tracing::warn! when a populated grid misses the key (signals a pixel→district convention bug). No more silently-wrong topology.
H2 Parity tripwire covered transit_hub only for Commission + Corporate Added Pioneer / Industrial / Military / Academic transit_hub rows — an accidental archetype-conditionalization of the cross-archetype override is now caught for every archetype.
Tyre (non-blocking) arrangement_pattern doc said "re-derived at L4" without saying where Clarified: the re-derivation call lives in the consumer (generate_quarter_skeleton), not in build_skeleton_work_item (which only threads inputs).

Also reconciled the branch's earlier export --stage post-checkout hook with main's superseding do-nothing version (b24a3596b) so the two merge cleanly. (Per pql-clide's review of the upstream feature request: do-nothing on same-HEAD is the minimal fix — no git-index side effects, no cross-branch leak.)

## Review findings addressed — `2d85f8ca3` All findings resolved. Gate re-run green: `cargo test` 1559+ lib + integration (0 failures), clippy `--all-targets -D warnings` clean, fmt clean. | # | Finding | Resolution | |---|---------|------------| | H1 | Silent `AlluvialPlain` morphology fallback | Now `tracing::debug!` when the district grid is empty (expected: params-missing body / early cascade) and `tracing::warn!` when a populated grid misses the key (signals a pixel→district convention bug). No more silently-wrong topology. | | H2 | Parity tripwire covered `transit_hub` only for Commission + Corporate | Added Pioneer / Industrial / Military / Academic `transit_hub` rows — an accidental archetype-conditionalization of the cross-archetype override is now caught for every archetype. | | Tyre (non-blocking) | `arrangement_pattern` doc said "re-derived at L4" without saying where | Clarified: the re-derivation call lives in the consumer (`generate_quarter_skeleton`), not in `build_skeleton_work_item` (which only threads inputs). | Also reconciled the branch's earlier `export --stage` post-checkout hook with main's superseding do-nothing version (`b24a3596b`) so the two merge cleanly. (Per pql-clide's review of the upstream feature request: do-nothing on same-HEAD is the minimal fix — no git-index side effects, no cross-branch leak.)
jpmschweitzer closed this pull request 2026-06-16 13:06:33 +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#168