diff --git a/.project/objectives/README.md b/.project/objectives/README.md index 51847bbf..2645a662 100644 --- a/.project/objectives/README.md +++ b/.project/objectives/README.md @@ -17,8 +17,8 @@ | **P0** | 44 | 0 | 0 | 0 | 0 | 0 | 44 | | **P1** | 88 | 0 | 0 | 0 | 0 | 1 | 89 | | **P2** | 130 | 0 | 0 | 0 | 0 | 1 | 131 | -| **P3 (oos)** | 29 | 0 | 0 | 0 | 0 | 29 | 58 | -| **total** | **291** | **0** | **0** | **0** | **0** | **31** | **322** | +| **P3 (oos)** | 29 | 0 | 3 | 0 | 3 | 29 | 64 | +| **total** | **291** | **0** | **3** | **0** | **3** | **31** | **328** | @@ -26,7 +26,7 @@ | Team Lead | Remaining | |---|---| -| — | 0 | +| [warcouncil](../team-leads/warcouncil.md) | 6 | diff --git a/.project/objectives/p3-19-player-ecology-feedback.md b/.project/objectives/p3-19-player-ecology-feedback.md new file mode 100644 index 00000000..77d26af7 --- /dev/null +++ b/.project/objectives/p3-19-player-ecology-feedback.md @@ -0,0 +1,52 @@ +--- +id: p3-19 +title: Player → ecology feedback — harvesting & hunting deplete live populations (over-harvest → extinction) +priority: p3 +status: missing +scope: game1 +owner: warcouncil +updated_at: 2026-06-25 +--- + +## Summary + +The living-world ecology sim (mc-ecology / mc-worldsim) ticks per played turn +(populations grow/collapse, migrate, emerge — see g2-08/g2-10, all done), and the +player can harvest flora (chop, `mc-city::harvest`) and kill fauna (combat / +lair clearing, `mc-combat::loot`). **But the two are decoupled:** harvesting and +hunting do NOT reduce the ecology `PopulationSlot::population` +(`mc-ecology/src/population.rs:9`) — fauna products are a read-only yield +(`fauna_product.rs`), and chopping flips terrain + grants a one-shot yield without +touching the live population. Only the sim's own tier dynamics cut populations +(`evolution.rs:1119 ecological_t5_reduces_fauna_populations`). + +Net: **you cannot over-harvest or over-hunt a species toward local extinction**, +and abundance doesn't respond to player pressure — which undercuts the +"living-world is the Game-1 USP" promise (the world should react to the player). + +## Acceptance + +- [ ] Killing fauna (combat / lair clear) decrements that species' live + `PopulationSlot::population` on/near the tile (scaled by group size killed). +- [ ] Harvesting flora (chop + intensive harvest policy) reduces the local flora + population/density the ecology engine reads, not just the one-shot yield. +- [ ] Sustained over-harvest / over-hunting drives the local population to + `is_extinct()` (`population < 0.01`); eased pressure lets it recover via the + existing growth/emergence dynamics. +- [ ] Logic lives in Rust (the coupling is in mc-ecology / mc-turn, not GDScript). +- [ ] Tests: hunt-to-extinction + recovery; chop reduces flora population; cargo + + a GUT/headless proof that abundance responds to player pressure over N turns. + +## Code sites + +- `mc-ecology/src/population.rs` (population mutation API — add a depletion method) +- `mc-ecology/src/fauna_product.rs` (currently read-only yield) +- `mc-combat/src/loot.rs` + the kill path (decrement on kill) +- `mc-city/src/harvest.rs` / `harvest_policy.rs` (chop / intensive → population) +- the per-turn coupling site (mc-turn or the worldsim tick driver) + +## Notes + +Verified 2026-06-25 (this is a genuine gap, not a wrongly-marked-done objective). +The ecology engine itself IS wired into the played turn (`turn_manager.gd:302` → +`mc-worldsim`); what's missing is the player→population coupling. diff --git a/.project/objectives/p3-20-weather-affects-scouting.md b/.project/objectives/p3-20-weather-affects-scouting.md new file mode 100644 index 00000000..b4a9aab1 --- /dev/null +++ b/.project/objectives/p3-20-weather-affects-scouting.md @@ -0,0 +1,42 @@ +--- +id: p3-20 +title: Weather affects scouting — vision/LoS penalty under storms, blizzards, dust +priority: p3 +status: missing +scope: game1 +owner: warcouncil +updated_at: 2026-06-25 +--- + +## Summary + +The runtime weather system (`mc-climate::weather`, six Game-1 types: Storm, Heat +Wave, Blizzard, Drought, Flood, Dust Storm) derives events per turn and applies a +`movement_penalty` to units, with real unit HP damage wired into the played turn +(`climate_effects.gd:125`). **But `WeatherEvent` carries no vision/line-of-sight +effect** — weather slows movement and hurts units, but does not reduce scouting +range. So storms/blizzards/dust have no effect on exploration, fog reveal, or +surprise — a missed gameplay lever for the weather system. + +## Acceptance + +- [ ] `mc-climate::weather::WeatherEvent` gains a `vision_penalty` (per type; + authored in the weather thresholds JSON, data-driven). +- [ ] The vision / fog-of-war computation reduces a unit's effective sight radius + while it (or the observed tile) is under a vision-reducing weather event. +- [ ] Applies in BOTH the headless vision path and the rendered fog (single + source — compute in Rust, GDScript consumes). +- [ ] Surfaced to the player (reduced reveal under weather is observable). +- [ ] Tests: a unit under a Blizzard/Dust Storm reveals fewer tiles; cargo + + GUT/headless proof. + +## Code sites + +- `mc-climate/src/weather.rs` (`WeatherEvent` + thresholds) +- the vision computation (`mc-vision` / `PlayerVision`; `pathfinder.gd::visible_hexes` + + the Rust vision projection) +- `public/resources/.../weather*.json` (per-type vision_penalty authoring) + +## Notes + +Verified 2026-06-25 — weather is `movement_penalty`-only; no vision field exists. diff --git a/.project/objectives/p3-21-weather-driven-migration.md b/.project/objectives/p3-21-weather-driven-migration.md new file mode 100644 index 00000000..50ffe908 --- /dev/null +++ b/.project/objectives/p3-21-weather-driven-migration.md @@ -0,0 +1,40 @@ +--- +id: p3-21 +title: Weather/climate-driven fauna & flora migration +priority: p3 +status: partial +scope: game1 +owner: warcouncil +updated_at: 2026-06-25 +--- + +## Summary + +Fauna/flora migration EXISTS and ticks per played turn +(`mc-ecology/src/biological.rs:59 MigrationPulse`, `engine.rs` per-turn migration; +g2-10 done), but it is **population-pressure / carrying-capacity driven only** — +it does not respond to **weather/climate**. Drought should push herds toward +water, blizzards/heat-waves should thin or relocate populations, blooms (already +climate-linked) should pull flora spread. Today the living world migrates by +crowding, not by climate — a partial realization of the weather↔ecology coupling. + +## Acceptance + +- [ ] Weather/climate events bias migration triggers and/or destinations + (e.g. drought raises migration-trigger chance + steers toward wetter tiles; + blizzard reduces local carrying capacity → emigration). +- [ ] Coupling is data-driven (per-weather migration modifiers authored in JSON, + not hardcoded). +- [ ] Logic in Rust (mc-ecology consumes the weather layer); GDScript only ticks. +- [ ] Tests: a drought shifts migration vs. a calm baseline (deterministic). + +## Code sites + +- `mc-ecology/src/biological.rs` (`MigrationPulse`, migration trigger/targeting) +- `mc-ecology/src/engine.rs` (per-turn migration step) +- `mc-climate::weather` (the weather layer to consume) + +## Notes + +Verified 2026-06-25 — migration is pop-pressure-driven; weather does not influence it. +Status `partial`: migration works, the weather-driving is the missing half. diff --git a/.project/objectives/p3-22-ai-builds-scouts.md b/.project/objectives/p3-22-ai-builds-scouts.md new file mode 100644 index 00000000..a18032c5 --- /dev/null +++ b/.project/objectives/p3-22-ai-builds-scouts.md @@ -0,0 +1,40 @@ +--- +id: p3-22 +title: AI builds dedicated scout units for exploration +priority: p3 +status: missing +scope: game1 +owner: warcouncil +updated_at: 2026-06-25 +--- + +## Summary + +The AI explores via frontier-seek with idle military units +(`mc-ai/src/tactical/movement.rs:892`) and gives scout-typed units a special +enemy-city patrol sweep — but it **never builds `dwarf_scout`**: the production +ladder (`mc-ai/src/tactical/production.rs:387`) queues founders, military, and +buildings, with workers as fallback, and no scout branch. Early-game discovery / +first-contact therefore relies on diverting combat units, which is slower and +weaker than a cheap dedicated scout. + +## Acceptance + +- [ ] `production.rs` queues a `dwarf_scout` early (e.g. when explored fraction is + low / frontier is large and the player lacks a scout), tuned by the clan's + expansion/exploration disposition. +- [ ] The built scout actually feeds exploration (frontier-seek / first-contact) + faster than the idle-military baseline. +- [ ] Logic in `mc-ai` (Rust). Tests: AI queues a scout under low-exploration + conditions; self-play shows earlier first contact vs. the no-scout baseline. + +## Code sites + +- `mc-ai/src/tactical/production.rs` (production priority ladder) +- `mc-ai/src/tactical/movement.rs` (scout sweep + frontier-seek already exist) +- `public/resources/units/dwarf_scout.json` + +## Notes + +Verified 2026-06-25. Minor relative to the ecology/weather gaps, but a real AI +quality gap (the AI under-explores in the opening). diff --git a/.project/objectives/p3-23-trade-richness-gold-strategic.md b/.project/objectives/p3-23-trade-richness-gold-strategic.md new file mode 100644 index 00000000..87bb960b --- /dev/null +++ b/.project/objectives/p3-23-trade-richness-gold-strategic.md @@ -0,0 +1,41 @@ +--- +id: p3-23 +title: Trade richness — gold & strategic-resource trades with opponents +priority: p3 +status: partial +scope: game1 +owner: warcouncil +updated_at: 2026-06-25 +--- + +## Summary + +Inter-player trade EXISTS but is **luxury-for-luxury swaps only** +(`mc-trade/src/lib.rs:139 evaluate_trades`; AI-driven, relation-gated, can't trade +your last copy). You **cannot trade gold or strategic resources** with opponents, +and the `tribute.rs` path is deferred to Game 2. That makes diplomacy/economy +thinner than the resource model implies (strategic resources gate units and are +tradeable in concept, but no exchange path exists). + +## Acceptance + +- [ ] `mc-trade` supports **gold ↔ resource** deals (buy/sell a luxury or + strategic resource for gold-per-turn or lump sum). +- [ ] `mc-trade` supports **strategic-resource** trades (e.g. iron for horses), + with the same "keep your last copy" protection. +- [ ] AI evaluates and proposes/accepts these deals (willingness + relation + + need, reusing the existing evaluation surface). +- [ ] Logic in Rust (`mc-trade`); GDScript only surfaces the deal UI. Tests: + gold-for-luxury and strategic-for-strategic deals evaluate + activate. + +## Code sites + +- `mc-trade/src/lib.rs` (`evaluate_trades`, `TradeAgreement`, `TradeLedger`) +- `mc-trade/src/tribute.rs` (deferred path — keep Game-2 gated) +- `public/resources/resources.json` (strategic `tradeable` flags) + +## Notes + +Verified 2026-06-25. Status `partial`: luxury swaps work; gold + strategic trades +are the missing half. (Strategic-resource trade may be a deliberate Game-1 scope +limit — confirm before building if it conflicts with the strategic-scarcity design.) diff --git a/.project/objectives/p3-24-rail1-economy-turn-logic-port.md b/.project/objectives/p3-24-rail1-economy-turn-logic-port.md new file mode 100644 index 00000000..1cb18c1f --- /dev/null +++ b/.project/objectives/p3-24-rail1-economy-turn-logic-port.md @@ -0,0 +1,51 @@ +--- +id: p3-24 +title: Rail-1 — port per-turn economy/happiness/climate glue logic from GDScript to Rust +priority: p3 +status: partial +scope: game1 +owner: warcouncil +updated_at: 2026-06-25 +--- + +## Summary + +The core economy/happiness/event MATH is in Rust (mc-economy, mc-happiness, +mc-city, mc-climate, mc-trade) — good Rail-1 health. **But real per-turn game +logic still lives in GDScript**, violating "GDScript is presentation only": + +- `economy.gd:66-72` computes gold *in GDScript* (building-gold sum, gold-per-pop, + gold-from-mines) before/around the `GdEconomy` call — not a thin wrapper. +- `happiness.gd` assembles + partially computes the happiness inputs (luxury map, + building-effect collection) in GDScript. +- `climate_effects.gd:125` mutates game state in GDScript (`unit.hp -= hp_loss`). +- `turn_processor.gd` / `turn_manager.gd` own the per-turn ORCHESTRATION (sequence + the Rust crates, dispatch events, apply results) in GDScript. + +This is the same class of debt as the AI port (p0-26, done) — but for the +economy/happiness/event/turn surface. + +## Acceptance + +- [ ] Gold income/expense aggregation (incl. per-pop, per-mine, building sums) + computed entirely in `mc-economy` from state; `economy.gd` only passes state + + applies the result. +- [ ] Happiness input assembly + computation fully in `mc-happiness`; `happiness.gd` + reduced to a thin bridge. +- [ ] Climate-effect damage application (unit HP loss) owned by Rust; GDScript + renders/animates only. +- [ ] (Stretch) per-turn orchestration moved behind a Rust turn driver so the + GDScript turn loop is a thin pump (overlaps the broader pathfinder/turn port). +- [ ] No regression: cargo + canonical GUT suite green. + +## Code sites + +- `src/game/engine/src/modules/empire/economy.gd` → `mc-economy` +- `src/game/engine/src/modules/empire/happiness.gd` → `mc-happiness` +- `src/game/engine/src/modules/climate/climate_effects.gd` → `mc-climate`/`mc-turn` +- `src/game/engine/src/modules/management/turn_processor.gd` (orchestration) + +## Notes + +Verified 2026-06-25. Phased / can land incrementally per surface. Related: p0-26 +(AI port, done), the standing pathfinder.gd port debt (p3-18 P5b note). diff --git a/public/games/age-of-dwarves/data/objectives.json b/public/games/age-of-dwarves/data/objectives.json index fe8a9766..e2f24aaf 100644 --- a/public/games/age-of-dwarves/data/objectives.json +++ b/public/games/age-of-dwarves/data/objectives.json @@ -1,13 +1,13 @@ { - "generated_at": "2026-06-25T11:28:51Z", + "generated_at": "2026-06-25T17:45:19Z", "totals": { + "partial": 3, "oos": 31, "done": 291, "stub": 0, - "partial": 0, "in_progress": 0, - "missing": 0, - "total": 322 + "missing": 3, + "total": 328 }, "objectives": [ { @@ -3229,6 +3229,66 @@ "owner": "warcouncil", "updated_at": "2026-06-25", "summary": "Land armies are permanently confined to their starting landmass: `mc-pathfinding`\ngates water purely by `UnitDomain` (`is_passable(\"ocean\", Land)` is hard-`false`),\nwith no tech override and no way to embark or be ferried. On maps where the two\ncapitals sit on separate landmasses (common — the start-balancer maximally\nseparates capitals on the 40×24 `duel` map, often ocean-separated), conquest by a\nland army is impossible. This is the Civ \"embarkation\" tech gap.\n\nThe scaffolding is **half-built** (existing tech debt): `mc_combat::siege::\nembarked_defence_penalty` (halves an embarked unit's defence — the Civ-V/VI\nvulnerability rule) and a `transport` keyword in `combat.json` (\"carry up to 2\nland units across water\", on `dwarf_fortress_ship`) both exist, but neither is\nwired into movement/pathfinding, so they are dead." + }, + { + "id": "p3-19", + "title": "Player → ecology feedback — harvesting & hunting deplete live populations (over-harvest → extinction)", + "priority": "p3", + "status": "missing", + "scope": "game1", + "owner": "warcouncil", + "updated_at": "2026-06-25", + "summary": "The living-world ecology sim (mc-ecology / mc-worldsim) ticks per played turn\n(populations grow/collapse, migrate, emerge — see g2-08/g2-10, all done), and the\nplayer can harvest flora (chop, `mc-city::harvest`) and kill fauna (combat /\nlair clearing, `mc-combat::loot`). **But the two are decoupled:** harvesting and\nhunting do NOT reduce the ecology `PopulationSlot::population`\n(`mc-ecology/src/population.rs:9`) — fauna products are a read-only yield\n(`fauna_product.rs`), and chopping flips terrain + grants a one-shot yield without\ntouching the live population. Only the sim's own tier dynamics cut populations\n(`evolution.rs:1119 ecological_t5_reduces_fauna_populations`).\n\nNet: **you cannot over-harvest or over-hunt a species toward local extinction**,\nand abundance doesn't respond to player pressure — which undercuts the\n\"living-world is the Game-1 USP\" promise (the world should react to the player)." + }, + { + "id": "p3-20", + "title": "Weather affects scouting — vision/LoS penalty under storms, blizzards, dust", + "priority": "p3", + "status": "missing", + "scope": "game1", + "owner": "warcouncil", + "updated_at": "2026-06-25", + "summary": "The runtime weather system (`mc-climate::weather`, six Game-1 types: Storm, Heat\nWave, Blizzard, Drought, Flood, Dust Storm) derives events per turn and applies a\n`movement_penalty` to units, with real unit HP damage wired into the played turn\n(`climate_effects.gd:125`). **But `WeatherEvent` carries no vision/line-of-sight\neffect** — weather slows movement and hurts units, but does not reduce scouting\nrange. So storms/blizzards/dust have no effect on exploration, fog reveal, or\nsurprise — a missed gameplay lever for the weather system." + }, + { + "id": "p3-21", + "title": "Weather/climate-driven fauna & flora migration", + "priority": "p3", + "status": "partial", + "scope": "game1", + "owner": "warcouncil", + "updated_at": "2026-06-25", + "summary": "Fauna/flora migration EXISTS and ticks per played turn\n(`mc-ecology/src/biological.rs:59 MigrationPulse`, `engine.rs` per-turn migration;\ng2-10 done), but it is **population-pressure / carrying-capacity driven only** —\nit does not respond to **weather/climate**. Drought should push herds toward\nwater, blizzards/heat-waves should thin or relocate populations, blooms (already\nclimate-linked) should pull flora spread. Today the living world migrates by\ncrowding, not by climate — a partial realization of the weather↔ecology coupling." + }, + { + "id": "p3-22", + "title": "AI builds dedicated scout units for exploration", + "priority": "p3", + "status": "missing", + "scope": "game1", + "owner": "warcouncil", + "updated_at": "2026-06-25", + "summary": "The AI explores via frontier-seek with idle military units\n(`mc-ai/src/tactical/movement.rs:892`) and gives scout-typed units a special\nenemy-city patrol sweep — but it **never builds `dwarf_scout`**: the production\nladder (`mc-ai/src/tactical/production.rs:387`) queues founders, military, and\nbuildings, with workers as fallback, and no scout branch. Early-game discovery /\nfirst-contact therefore relies on diverting combat units, which is slower and\nweaker than a cheap dedicated scout." + }, + { + "id": "p3-23", + "title": "Trade richness — gold & strategic-resource trades with opponents", + "priority": "p3", + "status": "partial", + "scope": "game1", + "owner": "warcouncil", + "updated_at": "2026-06-25", + "summary": "Inter-player trade EXISTS but is **luxury-for-luxury swaps only**\n(`mc-trade/src/lib.rs:139 evaluate_trades`; AI-driven, relation-gated, can't trade\nyour last copy). You **cannot trade gold or strategic resources** with opponents,\nand the `tribute.rs` path is deferred to Game 2. That makes diplomacy/economy\nthinner than the resource model implies (strategic resources gate units and are\ntradeable in concept, but no exchange path exists)." + }, + { + "id": "p3-24", + "title": "Rail-1 — port per-turn economy/happiness/climate glue logic from GDScript to Rust", + "priority": "p3", + "status": "partial", + "scope": "game1", + "owner": "warcouncil", + "updated_at": "2026-06-25", + "summary": "The core economy/happiness/event MATH is in Rust (mc-economy, mc-happiness,\nmc-city, mc-climate, mc-trade) — good Rail-1 health. **But real per-turn game\nlogic still lives in GDScript**, violating \"GDScript is presentation only\":\n\n- `economy.gd:66-72` computes gold *in GDScript* (building-gold sum, gold-per-pop,\n gold-from-mines) before/around the `GdEconomy` call — not a thin wrapper.\n- `happiness.gd` assembles + partially computes the happiness inputs (luxury map,\n building-effect collection) in GDScript.\n- `climate_effects.gd:125` mutates game state in GDScript (`unit.hp -= hp_loss`).\n- `turn_processor.gd` / `turn_manager.gd` own the per-turn ORCHESTRATION (sequence\n the Rust crates, dispatch events, apply results) in GDScript.\n\nThis is the same class of debt as the AI port (p0-26, done) — but for the\neconomy/happiness/event/turn surface." } ] }