feedback3 WD: the ammo bar drains out of the activation's real bullet total, 拉特兰 / 逃犯引渡手续 extras included (community report #35)

test/sim/feedback3-ammo-bar.test.js

@@ -0,0 +1,61 @@
+// test/sim/feedback3-ammo-bar.test.js — community report #35 (0.1.3): "拉特兰盟约叠层之后蕾缪安3技能锁定敌人消耗的子弹不会实时让技力条一格
+// 一格减少，而是最后突然消耗". 拉特兰's skillStart multiplies the bullets (×(1.05 + 0.015 × layers), floored) and 逃犯引渡手续 adds
+// one, but the snapshot drew ammoLeft / skill.ammo (the base 5) and the client clamps at 1: the bar stayed full until
+// fewer than 5 bullets were left. Now the draining bar is ammoLeft / ammoMax — the most bullets this activation held
+// (set at activation, raised by skillStart additions and addAmmo) — so it starts full and every bullet shortens it.
+
+import { test } from 'node:test';
+import assert from 'node:assert/strict';
+import { makeBattle, enemyRec, checkInvariants } from '../helpers/battleHarness.js';
+import { unitTuple } from '../../server/sim/snapshot.js';
+
+const DEFS = { enemies: { enemy_dummy: enemyRec({ key: 'enemy_dummy', hp: 1e7, speed: 0 }) } };
+const bar = (u, t) => { const [, , , , , sp, spMax] = unitTuple(u, t); return sp / spMax; };
+
+test('蕾缪安 S3 with 30 拉特兰 layers: the bar starts full and loses one bullet\'s share per lock (base 5 + extras)', () => {
+  const seen = [];
+  const h = makeBattle({
+    defs: DEFS, timeLimit: 200, seed: 7,
+    units: [{ chessId: 'chess_char_6_01_a', row: 10, col: 3 }],
+    enemies: [{ key: 'enemy_dummy', pos: [10, 5] }],
+    bonds: { lateranoShip: { count: 3, active: true, tier: 1, layers: 30 } },
+    setup(b) {
+      b.on('ammoUsed', ({ unit, left, skill }) => seen.push({ left, max: skill.ammoMax, bar: bar(unit, b.time) }), { priority: -999 });
+    },
+  });
+  const u = h.unit('chess_char_6_01_a');
+  assert.equal(u.skill.ammo, 5, 'S3: 5 bullets (attack@trigger_time)');
+  assert.ok(h.runUntil(() => u.skill.active, 80));
+  assert.ok(h.runUntil(() => !u.skill.active, 30));
+  assert.ok(seen.length >= 7, `${seen.length} bullets: 5 × 1.5 floored, + 逃犯引渡手续`);
+  const full = seen[0].left + 1;
+  for (const s of seen) assert.equal(s.max, full, 'the full mark is the activation\'s real total');
+  seen.forEach((s, i) => assert.ok(Math.abs(s.bar - (full - 1 - i) / full) < 0.01, `bullet ${i + 1}: bar ${s.bar.toFixed(3)} = ${full - 1 - i}/${full}`));
+  assert.ok(seen[0].bar < 1 - 0.1, 'the first lock already shortens the bar (it used to stay full until < 5 were left)');
+  assert.equal(u.skill.ammoMax, 0, 'reset when the skill ends');
+  checkInvariants(h.b);
+});
+
+test('the full mark follows bullets added mid-skill only above it; without extras the bar is unchanged (n / base)', () => {
+  const h = makeBattle({
+    defs: DEFS, timeLimit: 200, seed: 7,
+    units: [{ chessId: 'chess_char_6_01_a', row: 10, col: 3 }],
+    enemies: [{ key: 'enemy_dummy', pos: [10, 5] }],
+  });
+  const u = h.unit('chess_char_6_01_a');
+  h.step(1);
+  u.skill.activate('test', { free: true });
+  const sk = u.skill;
+  assert.equal(sk.ammoMax, sk.ammoLeft);
+  const base = sk.ammoMax;
+  assert.ok(Math.abs(bar(u, h.b.time) - 1) < 1e-9, 'full at the start');
+  sk.ammoLeft -= 2;                                 // two locks
+  assert.ok(Math.abs(bar(u, h.b.time) - (base - 2) / base) < 0.01);
+  sk.addAmmo(1);                                    // a refill under the mark: the mark stays
+  assert.equal(sk.ammoMax, base);
+  sk.addAmmo(3);                                    // above it: the mark rises, the bar is full again
+  assert.equal(sk.ammoMax, base + 2);
+  assert.ok(Math.abs(bar(u, h.b.time) - 1) < 1e-9);
+  sk.end('test');
+  assert.equal(sk.ammoMax, 0);
+});