rimlike

megaproxy/rimlike

Fork 0

Commit graph

Author	SHA1	Message	Date
megaproxy	e7a2407af2	Job: target_node claim so pawns don't cluster on the same target Visible bug: with 1 wall ghost queued, all 3 pawns picked the same site; 2 stood idle while 1 built. Same shape would affect chop/mine/haul/etc. Design: Job carries an untyped target_node ref (the tree/rock/build-site/ crop/item/patient/workbench/etc the job is acting on). Job.is_target_taken_ by_other(target, excluding_pawn) does an O(pawns) scan of live job state to ask 'is anyone else already working this?'. Each WorkProvider's find_best_ for() now skips claimed targets in its scan and sets j.target_node before returning. No per-entity claim state, no .claim()/.release() bookkeeping, no save-format change — target_node is intentionally not serialized because pawns re-decide and re-bind naturally after load. Providers updated: construction / chop / mine / plant (harvest path) / hauling (item AND corpse) / cleaning (target is Vector2i tile not Node; field is untyped, doc'd) / doctor / crafting (workbench). Not touched: rest (everyone shares the rest tile, that's fine), eat / sleep (food and beds have their own availability gates; flagged as a followup if multi-pawn food contention surfaces). Verified MCP runtime: fresh boot, 3 pawns picked 3 distinct wall sites (44,28)/(45,28)/(44,27) with distinct target_node refs. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-05-12 13:45:44 +01:00
megaproxy	9cf9b7dbfd	Phase 13: Rooms + Auto-roof + Beauty + Dirtiness + Cleaning Three-agent fan-out — Opus pre-wrote Room class, World.rooms/room_at_tile/is_indoor, 4 EventBus signals before dispatch so the slices ran fully parallel. DECISION: Big-room UX = bump auto-roof cap to 16, banner above. Cabin (24 tiles) intentionally exceeds cap to exercise the warning path; a 5×5 test shed (9 interior tiles) was added to exercise the roof path. Room detection (Agent A): - scenes/world/room.gd — class_name Room, tiles/bounds/is_under_roof, contains_tile() bounds-then-list-checked, recompute_bounds() - scenes/world/room_detector.gd — class_name RoomDetector, BFS 4-dir from floor/door tiles, walls/terrain as boundary, doors counted as room interior. Detects up to 4× cap; auto-roofs only ≤16. - World.mark_wall_tile/mark_floor_tile/mark_door_tile hook BFS recompute - Door._complete() now erases wall-layer stamp + registers door tile - Designation.TOOL_NO_ROOF paint mode wired (UI button deferred Phase 17) - EventBus.room_changed / room_too_large signals Indoor/Shelter (Agent B): - Pawn._is_sheltered() rerouted: World.is_indoor() first, floor-proxy fallback - IndoorTintOverlay Node2D — _draw fills roofed-room tiles at α=0.10 warm - Crop._on_sim_tick skips stage advance when World.is_indoor(tile) Beauty + Dirtiness + Cleaning + Room thoughts (Agent C): - BeautySystem sparse map, linear falloff radius=3, Quality multiplier (SHODDY 0.5 → LEGENDARY 2.5). Base: Bed +2, Workbench +1, Torch +3, Hearth +4 - DirtinessSystem 0-100, tier crossings (clean<25/dirty<60/filthy≥60) emit tile_dirtiness_changed. bump/bump_clean/bump_pawn_traffic API - CleaningProvider priority=2, KIND_CLEAN toil, 2.5 dirt/tick for ~40 ticks - Bed/Torch/Workbench _complete() now register with BeautySystem - 7 room mood thoughts: clean_room (+2), dirty_room (-3), filthy_room (-6), beautiful_room (+4), ugly_room (-3), slept_in_room (+3 EVENT, wires Ph 17), ate_without_table (-3 EVENT, wires Ph 17) - Pawn._sync_room_thoughts called from _process_thoughts after cold block, defensive against null rooms/systems Integration recovery (Opus): - Agent C's BeautySystem/DirtinessSystem/CleaningProvider/IndoorTintOverlay instantiation in world.gd never landed (only field declarations + entity hooks survived). Added preloads + runtime add_child + autoload bindings + CleaningProvider registration + furniture pre-seed in _ready - Added _prestamp_test_shed_for_room_detector with _spawn_complete_wall/floor helpers so a 5×5 visible shed exercises the auto-roof path at boot MCP runtime verified: - Rooms: cabin Room#2 size=24 roofed=false (room_too_large fires), shed Room#3 size=9 roofed=true (auto-roof active) - beauty_map size=50 around prebuilt furniture; bed at (47,24) beauty=4.0 - Bram teleported to (36, 25) in shed → indoor=true, sheltered=true, thoughts=[clean_room +2], mood=52.0 - Screenshot: shed walls + brown floor visible; cabin warmly torch-lit; Spring 1/12 indicator; Day 1 07:52 Delegation: 3× gdscript-refactor (Sonnet) agents in parallel; integration recovery + MCP verify on Opus. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>	2026-05-11 17:19:23 +01:00

Author

SHA1

Message

Date

megaproxy

e7a2407af2

Job: target_node claim so pawns don't cluster on the same target

Visible bug: with 1 wall ghost queued, all 3 pawns picked the same site;
2 stood idle while 1 built. Same shape would affect chop/mine/haul/etc.

Design: Job carries an untyped target_node ref (the tree/rock/build-site/
crop/item/patient/workbench/etc the job is acting on). Job.is_target_taken_
by_other(target, excluding_pawn) does an O(pawns) scan of live job state to
ask 'is anyone else already working this?'. Each WorkProvider's find_best_
for() now skips claimed targets in its scan and sets j.target_node before
returning. No per-entity claim state, no .claim()/.release() bookkeeping,
no save-format change — target_node is intentionally not serialized
because pawns re-decide and re-bind naturally after load.

Providers updated: construction / chop / mine / plant (harvest path) /
hauling (item AND corpse) / cleaning (target is Vector2i tile not Node;
field is untyped, doc'd) / doctor / crafting (workbench).

Not touched: rest (everyone shares the rest tile, that's fine), eat /
sleep (food and beds have their own availability gates; flagged as a
followup if multi-pawn food contention surfaces).

Verified MCP runtime: fresh boot, 3 pawns picked 3 distinct wall sites
(44,28)/(45,28)/(44,27) with distinct target_node refs.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-12 13:45:44 +01:00

megaproxy

9cf9b7dbfd

Phase 13: Rooms + Auto-roof + Beauty + Dirtiness + Cleaning

Three-agent fan-out — Opus pre-wrote Room class, World.rooms/room_at_tile/is_indoor,
4 EventBus signals before dispatch so the slices ran fully parallel.

DECISION: Big-room UX = bump auto-roof cap to 16, banner above. Cabin
(24 tiles) intentionally exceeds cap to exercise the warning path; a
5×5 test shed (9 interior tiles) was added to exercise the roof path.

Room detection (Agent A):
- scenes/world/room.gd — class_name Room, tiles/bounds/is_under_roof,
  contains_tile() bounds-then-list-checked, recompute_bounds()
- scenes/world/room_detector.gd — class_name RoomDetector, BFS 4-dir
  from floor/door tiles, walls/terrain as boundary, doors counted as
  room interior. Detects up to 4× cap; auto-roofs only ≤16.
- World.mark_wall_tile/mark_floor_tile/mark_door_tile hook BFS recompute
- Door._complete() now erases wall-layer stamp + registers door tile
- Designation.TOOL_NO_ROOF paint mode wired (UI button deferred Phase 17)
- EventBus.room_changed / room_too_large signals

Indoor/Shelter (Agent B):
- Pawn._is_sheltered() rerouted: World.is_indoor() first, floor-proxy fallback
- IndoorTintOverlay Node2D — _draw fills roofed-room tiles at α=0.10 warm
- Crop._on_sim_tick skips stage advance when World.is_indoor(tile)

Beauty + Dirtiness + Cleaning + Room thoughts (Agent C):
- BeautySystem sparse map, linear falloff radius=3, Quality multiplier
  (SHODDY 0.5 → LEGENDARY 2.5). Base: Bed +2, Workbench +1, Torch +3, Hearth +4
- DirtinessSystem 0-100, tier crossings (clean<25/dirty<60/filthy≥60)
  emit tile_dirtiness_changed. bump/bump_clean/bump_pawn_traffic API
- CleaningProvider priority=2, KIND_CLEAN toil, 2.5 dirt/tick for ~40 ticks
- Bed/Torch/Workbench _complete() now register with BeautySystem
- 7 room mood thoughts: clean_room (+2), dirty_room (-3), filthy_room (-6),
  beautiful_room (+4), ugly_room (-3), slept_in_room (+3 EVENT, wires Ph 17),
  ate_without_table (-3 EVENT, wires Ph 17)
- Pawn._sync_room_thoughts called from _process_thoughts after cold block,
  defensive against null rooms/systems

Integration recovery (Opus):
- Agent C's BeautySystem/DirtinessSystem/CleaningProvider/IndoorTintOverlay
  instantiation in world.gd never landed (only field declarations + entity
  hooks survived). Added preloads + runtime add_child + autoload bindings +
  CleaningProvider registration + furniture pre-seed in _ready
- Added _prestamp_test_shed_for_room_detector with _spawn_complete_wall/floor
  helpers so a 5×5 visible shed exercises the auto-roof path at boot

MCP runtime verified:
- Rooms: cabin Room#2 size=24 roofed=false (room_too_large fires),
  shed Room#3 size=9 roofed=true (auto-roof active)
- beauty_map size=50 around prebuilt furniture; bed at (47,24) beauty=4.0
- Bram teleported to (36, 25) in shed → indoor=true, sheltered=true,
  thoughts=[clean_room +2], mood=52.0
- Screenshot: shed walls + brown floor visible; cabin warmly torch-lit;
  Spring 1/12 indicator; Day 1 07:52

Delegation: 3× gdscript-refactor (Sonnet) agents in parallel;
integration recovery + MCP verify on Opus.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

2026-05-11 17:19:23 +01:00

2 commits