mirror of
https://github.com/kuhyx/testsAndMisc.git
synced 2026-11-12 08:51:01 +01:00
Krzysztof kuhy Rudnicki
2d66bfdd61
Ambush drops a unit beside the survivor, inside the ring its kiting AI would otherwise keep clear. Frenzy drives the whole horde faster and harder for six seconds. Rift pins every spawn to one edge for ten, so you can mass a flank instead of scattering. This reverses the README's old promise that placement is never yours to choose. The interesting decision turned out to be what a placement is worth, not being denied one — so placement is purchasable, never free: a premium for Ambush, a cooldown and a commitment window for Rift. A shared spendPower gate covers cooldown and affordability for all three, so the next power costs no new branches. EDGES became a Record keyed by a finite union rather than an array: under noUncheckedIndexedAccess an array index is possibly-undefined and would force an unreachable fallback. The unpinned spawn path still draws the same RNG values in the same order, verified against the old implementation across 50 seeds — existing seeded replays are unchanged. 145 tests, 100% coverage. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VFyXexNjMQG9V7K8de17fc
48 lines
3.5 KiB
JSON
48 lines
3.5 KiB
JSON
{
|
|
"intent": "Add three director powers - ambush, frenzy and rift - giving the player priced control over enemy placement and a burst tempo lever.",
|
|
"scope": [
|
|
"reverse_survivors/src/core/types.ts (PowerKind, costs/cooldowns, EdgeId, DirectorState fields, new DirectorAction variants)",
|
|
"reverse_survivors/src/core/director.ts (spendPower gate, ambush/frenzy/rift, rift-aware edgeSpawnPos)",
|
|
"reverse_survivors/src/core/sim.ts (frenzy speed and damage), snapshot.ts, src/ui/SpawnBar.tsx, src/ui/Hud.tsx, src/styles.css",
|
|
"Non-goal: difficulty tiers ship as their own commit"
|
|
],
|
|
"changes": [
|
|
"A shared spendPower gate handles cooldown and affordability for all three powers, so each additional power costs no new branches in the dispatcher.",
|
|
"EDGES converted from an array to Record<EdgeId, EdgeFn>: under noUncheckedIndexedAccess an array index would be possibly-undefined and force an unreachable fallback branch, whereas a finite-union Record stays non-optional.",
|
|
"Ambush placement uses a golden-angle step so successive summons never reuse a lane. An earlier 2*PI/7 formulation was caught in review: it silently cycled every 7 summons, with index 7 landing byte-identically on index 0.",
|
|
"Frenzy damage derives from one shared frenzyDamageMul helper. Review caught that an earlier version multiplied contact damage only, leaving enemy projectiles and bomber blasts unscaled - which contradicted the README and left the bomber, whose contactDamage is 0, entirely unaffected.",
|
|
"MOVE_FNS gained a speed parameter so frenzy (and later difficulty) multiply in one place rather than at three call sites.",
|
|
"The README's former promise that enemy placement is never the player's to choose is deliberately reversed here and documented as such."
|
|
],
|
|
"verification": [
|
|
{
|
|
"command": "npm run check (typecheck + eslint --max-warnings 0 + vitest --coverage)",
|
|
"result": "pass",
|
|
"evidence": "145 tests passed. Coverage 100% on all four metrics: statements 516/516, branches 153/153, functions 96/96, lines 499/499."
|
|
},
|
|
{
|
|
"command": "ambush lane probe via vite-node",
|
|
"result": "pass",
|
|
"evidence": "24 consecutive ambushes produced 24 distinct spawn positions. Before the golden-angle fix the same probe produced only 7, with index 7 identical to index 0 at (610.0000, 300.0000)."
|
|
},
|
|
{
|
|
"command": "frenzy damage-channel probe via vite-node",
|
|
"result": "pass",
|
|
"evidence": "Ratios with frenzy on vs off, measured over 360 ticks: warden contact 1.300, artillery projectile 1.300, stalker projectile 1.300, bomber blast 1.300. Before the fix the last three measured 1.000."
|
|
},
|
|
{
|
|
"command": "RNG-identity check against the pre-change implementation",
|
|
"result": "pass",
|
|
"evidence": "The old array-based pick was reimplemented from git HEAD and compared across 50 seeds x 20 calls on both the returned position and the final RNG state: 0 mismatches. Existing seeded replays are unchanged."
|
|
}
|
|
],
|
|
"risks": [
|
|
"Powers are gated by cooldown rather than energy alone, because late-game income reaches 24/s and an energy-only gate would degenerate into spamming the highest-value button.",
|
|
"Ambush lets the director bypass the survivor's kiting AI, which is the intended counter to its late-game power curve but is the most likely thing to need retuning."
|
|
],
|
|
"rollback": [
|
|
"git revert the commit; no persisted state or data migration is involved.",
|
|
"After rollback run `npm run check` in reverse_survivors/ and confirm the suite passes at 100% coverage."
|
|
]
|
|
}
|