sganggs 3dd2213e47 0.2.5-dev: merge claude/036-sim ...

# Conflicts:
#	CHANGELOG.md
#	docs/history/0.2.5.md
#	test/golden/matches.json

2026-10-12 15:50:42 +08:00

..

battle.test.js

Operator battle voices: the official combat lines play by voice slot during battles, one at a time (#73)

2026-10-06 01:44:46 +08:00

bossfield.browser.test.js

0.2.4-dev: round-29 community fixes and the official shared card pool

2026-10-11 01:58:52 +08:00

bosshp_stacking.test.js

Playtest #6 follow-up (v2.5.1): leader HP review — 直接乘算, one debuff instance, 999 layer cap, 限伤; elite to the board; 假想敌：胄 kit

2026-10-02 04:19:28 +08:00

combat.test.js

feedback3: merge WG — nest enemies pulse instead of attacking, children carry no bounty, a 隐匿 blocker is skipped by areas, stacked cold freezes for the longer duration

2026-10-04 21:25:40 +08:00

core.test.js

Playtest #6 fixes: summons placed by hand, boss pool stall, bounties, 联防 leak counter, distinct offers, 机变 card text, phone buy tap, skill triggers, blocking, push force, 孤立, enemies, element bar

2026-10-01 19:08:23 +08:00

detmath.test.js

detmath powi: a zero base keeps Math.pow's exact results ((-0)^1 = -0, (-0)^-1 = -Infinity)

2026-10-11 21:09:25 +08:00

displace-fx.test.js

Preserve enemy facing through push and pull slides (#418, #427)

2026-10-09 17:16:49 +08:00

docs-examples.test.js

Playtest #6 fixes: summons placed by hand, boss pool stall, bounties, 联防 leak counter, distinct offers, 机变 card text, phone buy tap, skill triggers, blocking, push force, 孤立, enemies, element bar

2026-10-01 19:08:23 +08:00

downelem.test.js

feedback2 WK: an operator forced off the field (史尔特尔's 余烬, 耀骑士临光 S2, 骑士戒律 + 竞技旗, 伊内丝 S3, a 商人 that cannot pay) lies down on its tile and redeploys there like a knocked-out one, still no kill (PRTS 卫戍协议/帮助; GitHub #60)

2026-10-04 02:55:48 +08:00

engine-requests.test.js

feedback2 WK: an operator forced off the field (史尔特尔's 余烬, 耀骑士临光 S2, 骑士戒律 + 竞技旗, 伊内丝 S3, a 商人 that cannot pay) lies down on its tile and redeploys there like a knocked-out one, still no kill (PRTS 卫戍协议/帮助; GitHub #60)

2026-10-04 02:55:48 +08:00

facing-invariance.test.js

feedback6 R21A: a 突袭 member may land on a teammate's half when both halves of the field are taken — the two-helper 联防 field and the Final Assault / Hidden Core pair field (the owner's decision of 2026-10-07): Battle.onFieldBoard (onOwnBoard of any player of the battle) read by raidTile only; a lone 联防 helper's empty half, a solo boss field's other half and the hand / 临时整备区 rows stay closed, every other landing rule kept, 乌尔比安's S3 【移动】 keeps onOwnBoard; golden: 8 moved, each parting from the parent at a member's first landing on its teammate's half (checked tick by tick with the parent instrumented) — fields boss-boss_2 / 3 / 4 / 5 / 6-pair and hidden-boss_8 / 10-pair (the right player's 缄默德克萨斯 / 伊内丝 / 仇白 / 耀骑士临光 on the left player's half, 1–7 landings per battle, the first 0.3–39.8 s in), matches coop4-ABYSS-7 (round 6's two-helper 联防: ai_2's 斯卡蒂 on ai_1's half at 52 s; only ai_1 / ai_2 dmgDealt move, 8263 between them); unite-2 unchanged (its two landings stay on their own half)

2026-10-07 17:09:39 +08:00

facing.test.js

feedback5 WE2: remove the kits/tier1-6.js re-export shims and shared/standIn.js diyRecord (#5)

2026-10-06 16:07:49 +08:00

feedback1-followups.test.js

feedback1 integrate: follow-ups from the workstream notes (official rules)

2026-10-03 07:57:28 +08:00

feedback1-integrate.test.js

feedback1 integrate: test that 维娜·维多利亚's S3 lions never land in 战场#08's 深水区 (WH x WC)

2026-10-03 06:54:34 +08:00

feedback1-residuals.test.js

feedback1 integrate: 坚固维式重锤's lock belongs to the deployment - a lent hammer and in-place revives follow it (QA)

2026-10-03 17:47:07 +08:00

feedback1-tank-triggers.test.js

feedback5 WF: ACTIVE_RANGE also replaces the SEARCH row and covers 深巡 S2 — a skill whose running range strictly contains the own one casts with an enemy inside it (the owner's decisions of 2026-10-05)

2026-10-05 18:45:13 +08:00

feedback1-transform.test.js

feedback5 WE1: 「同时攻击阻挡的所有敌人」 takes up to the block count of targets from the range, blocked first (the queued centurion question; Mon3tr S3)

2026-10-06 16:28:31 +08:00

feedback1-water.test.js

feedback1 WC: tactician summons never left out of range with a full hand; source fixes

2026-10-03 02:46:07 +08:00

feedback1b-arcane.test.js

feedback1b WG: 奥术 re-measured; bonds the mode never activates read 本局禁用

2026-10-03 02:31:35 +08:00

feedback1b-displacement.test.js

feedback7 R22-UNBAL: a pushed / pulled enemy is held in the official 失衡 (UNBALANCE) state — no walking, no normal attack — for the push's 位移时间 of its 受力等级 or the pull's force window; a 静态刚体 hit by a force gets the 0.1 s floor; the client's slide takes the same game time (idea from PR #392 by @xcdoge, re-implemented from the sources); golden: 69 scenarios move, 35 with kills / leaks / deaths / boss HP (below)

2026-10-08 21:54:06 +08:00

feedback1d-bombd.test.js

feedback1d WN: a stun before 暴鸰's release interrupts the cast; the release lands on tick 8; SIM.md timeline fixed

2026-10-03 05:55:29 +08:00

feedback1d-push.test.js

feedback1d WM: cite the client push templates for 野鬃 / 琳琅诗怀雅

2026-10-03 04:31:36 +08:00

feedback1e-aoe.test.js

feedback1e WR: real-wave test counts the enemies each attack damages; AoE source wording

2026-10-03 06:12:34 +08:00

feedback1e-healers.test.js

feedback7 R22-FIELDS: the Hidden Core 铳's 碎铳之簧 chase the casting gun for the whole 5 s of 末日布道 instead of stopping on arrival; each gun keeps its data timer, the later call takes the springs (PR #347 by @CXUtk, its pair stagger and lock not taken); golden: hidden-boss_9-pair, hidden-boss_9-solo, coop2-ABYSS-10-boosted

2026-10-08 19:13:21 +08:00

feedback1e-skillrange.test.js

feedback5 WC: a MANUAL skill whose running attack range strictly contains the own range casts as soon as an enemy is inside it (ACTIVE_RANGE, the owner's rule of 2026-10-05)

2026-10-05 17:44:33 +08:00

feedback1f-downed-tile.test.js

feedback5: merge CB2 — community reports (healers / summons): 刺玫's S2 heals the most injured ally again (the counter's heal alone goes to her protégé), 琳琅诗怀雅 heals herself and her 破财消灾 cost restarts at 5 DP with every deployment, 赫默's S2 casts at full SP so the drone takes the field at once, the countdown summons (医疗探机, 诅咒娃娃, 海嗣 and four 自选 ones) take no damage, are never healed and show the life left on their bar, stacked ground areas stay readable; §25.16 after CB5's §25.19 (renumbered with the later merges); golden: vs feedback5 CB2's own rules — roster-001/002/003/005/008/016/018/027/035/037/045/046/049, bond-deputShip-low/high, bond-suntShip-high, bond-victoriaShip-low/high, bond-visiShip-low/high, boss-boss_1/3/5/7-pair, hidden-boss_9-pair, unite-1, matches solo-FUNNY-2, coop2-HARD-4, coop3-ABYSS-5, coop4-ABYSS-7, coop2-NORMAL-8-serverrun, coop2-ABYSS-10-boosted, coop2-NORMAL-14-standins; vs CB2 exactly CB5's 18 bot matches

2026-10-07 03:10:22 +08:00

feedback1f-fence.test.js

feedback5 WF: ACTIVE_RANGE also replaces the SEARCH row and covers 深巡 S2 — a skill whose running range strictly contains the own one casts with an enemy inside it (the owner's decisions of 2026-10-05)

2026-10-05 18:45:13 +08:00

feedback1f-liftoff.test.js

feedback3: the airborne-蒂比 烹泉 test places the 烹泉 within its official 1.25 blast of both units and reads the per-blast 【烹泉减益】 layers (left behind by the W3B merge)

2026-10-04 19:53:24 +08:00

feedback2-doll.test.js

feedback3 WF: a 傀儡师 leaves every switch at full HP (PRTS: the HP is reset during the switch, 不死 throughout) — a second 阿戈尔 devour mark inside the 1-s switch no longer leaves 归溟幽灵鲨's 替身 at 1 HP (community report #2)

2026-10-04 19:27:34 +08:00

feedback2-enemy-attack-stand.test.js

feedback5 WE1: enemy attacks strike at their damage frame; a short stun before it cuts the swing (GitHub #187 point 1, #170)

2026-10-06 15:21:24 +08:00

feedback2-forced-exit-down.test.js

Deploy-timed skills run the normal duration lifecycle: one charge on deployment, a duration bar, a real end, again on the next deployment (#109)

2026-10-05 22:36:29 +08:00

feedback2-stealth-splash.test.js

feedback3 WG: a 隐匿 blocker is hit only by the blocked enemy's attacks, not by its areas, zones or auras (GitHub #97)

2026-10-04 20:33:35 +08:00

feedback2-surtr-stealth.test.js

feedback7 R23-TICK: an attack interval or a time-SP cost of a whole number of frames takes exactly that many — 1 s is 30 frames, not 31 (PR #402 by @Cloudnyco, ported onto 0.2.2); golden: every family moves (roster, bonds, fields, matches, standins, diy — regenerated in the next commit)

2026-10-09 01:37:32 +08:00

feedback2-whitw2-yu.test.js

feedback2 WH: 余 S2 厚礼上宾 casts with an enemy on its x-1 instead of when hit (the owner's deliberate deviation from the 重装 TAKE_DAMAGE row: TRIGGER_DEVIATIONS SKILL_RANGE, data regenerated, only the two 余 records change)

2026-10-04 00:39:54 +08:00

feedback3-ammo-bar.test.js

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)

2026-10-04 17:17:17 +08:00

feedback3-bounty-children.test.js

feedback3 WG: a split or summoned child does not carry the kill bounty, and 联防 pays a bounty card only when the leaked enemy is that card's enemy

2026-10-04 20:15:12 +08:00

feedback3-cold-freeze.test.js

feedback5 CB1: a pair of colds on an enemy becomes the freeze, no cold is left — the 谢拉格 wind no longer freezes for good from 51 layers on (community reports 「谢拉格盟约冰冻时间没有随层数正确成长」 and 「…被在无法被任何干员攻击到的地方永控」, items 7 and 8)

2026-10-07 01:39:50 +08:00

feedback3-stunsp.test.js

feedback3 WA: a stunned / frozen operator keeps its natural SP recovery (only 阻回 stops it) — 洛洛's S2 self-stun no longer freezes her SP

2026-10-04 17:19:41 +08:00

feedback3-swire-s3.test.js

feedback3 WI: 琳琅诗怀雅 S3 shoots at 10 coins only when an enemy is in her attack range

2026-10-04 21:11:41 +08:00

feedback3-ulpia-move.test.js

feedback4: 阿戈尔's devoured base ATK is a 最终加算, and a marker knocked out during the devour still resolves its own marks (GitHub #165 points 2 and 3, PR #176 by @SpiritedAwayCN; the owner's decision of 2026-10-06)

2026-10-06 14:39:32 +08:00

feedback4-restart-in-end.test.js

feedback4: a skill restarted inside its own end owns its events — Skill.end emits no trailing skillEnd / client 'skill off' when onEnd re-activated it (耀骑士临光 S2 retreats at its duration end and 不屈 redeploys her inside that call: the trailing skillEnd cleared 骑士戒律 for the whole second cast, a PR #109 regression); 伊内丝's 影哨 death hooks run before 不屈's redeploy (priority 20 over 10), so the next deployment's S3 recalls the sentry she left (0.1.4 release review); test/sim/feedback4-restart-in-end.test.js, each case fails without its fix; golden unchanged

2026-10-06 00:12:26 +08:00

feedback4-sleep-unblock.test.js

feedback4: a sleeping enemy cannot be blocked and takes no block slot — 沉睡 = 无法行动+无敌+不可阻挡 (GitHub #140; the owner's decision of 2026-10-06)

2026-10-06 16:57:55 +08:00

feedback5-active-range.test.js

feedback5 WF: ACTIVE_RANGE also replaces the SEARCH row and covers 深巡 S2 — a skill whose running range strictly contains the own one casts with an enemy inside it (the owner's decisions of 2026-10-05)

2026-10-05 18:45:13 +08:00

feedback5-attack-clip-rest.test.js

feedback5 R19E: an enemy finishes its attack clip after its block ends — the 鸭爵 swap's 流泪小子 no longer walks on the tick its strike stuns its blocker (community report of 2026-10-07 「哭狗瞬间隐匿没有后摇」, GitHub #214 by @FrogThai); golden: 216 scenarios move

2026-10-07 12:56:12 +08:00

feedback5-block-targets.test.js

feedback5 WE1: 「同时攻击阻挡的所有敌人」 takes up to the block count of targets from the range, blocked first (the queued centurion question; Mon3tr S3)

2026-10-06 16:28:31 +08:00

feedback5-bounty-payee.test.js

feedback6 R21B: 联防 plays on the round's battlefield again — its terrain, crates, water, devices and runes, as in 0.1.4 (the owner's decision of 2026-10-07, 「官服保留地形，改回去」; GitHub #244). 0.2.0's escaped-level map (8a436249, §25.6.4) is withdrawn: that map is the placeholder grid all 39 wave template levels carry, tile for tile; only the 11 stage levels have terrain. match/unitePhase.js _uniteOpts fields this.stageId in both combat modes, unite.js uniteStageId is removed, the client draws the 联防 on the board's stage again (screens/game.js: 0.1.4's ownStage / baseStage, no map switch, nothing to restore on leaving), the dev mock likewise. Kept from 0.2.0: own-board 突袭 / 乌尔比安 S3 landings (onOwnBoard, raidTile untouched), the bounty payee, the 联防 BGM, the SETTLE result box. data/stages.json keeps the two kind 'unite' records (no field uses them; engine tests use them as a plain two-halves road); tools/build-data.mjs comments only, an offline rebuild reproduces data/ byte for byte. The 联防 battle input is byte-identical to v0.1.4's in 20 cases (5 stages × 1 / 2 helpers × both combat modes); GitHub #41 item 3 (野鬃 pushing walkers back across the gate onto the other half) is back with v0.1.4's numbers, none off the map. Tests: test/match/feedback5-unite-map.test.js rewritten (6, all fail on the parent), test/e2e/coop.e2e.mjs checks the 联防 field's drawn stage; docs: DATA §12, META §4, PLAYING §5, design/geometry-time, I18N, the §25.6.4 pointer; golden: fields unite-1 (escaped_single → act1autochess_m02) and unite-2 (escaped_multi → act1autochess_m07), the stages of 0.1.4, and the co-op matches whose 联防 rounds come out differently on the round's map — coop2-NORMAL-3, coop2-HARD-4, coop3-ABYSS-5, coop4-ABYSS-7, coop2-NORMAL-8-serverrun, coop2-NORMAL-14-standins (now a victory)

2026-10-07 17:40:03 +08:00

feedback5-catapult-smog.test.js

feedback5 CB4: 自制投石机's targeting ignores its block (索敌不受阻挡影响) — blocked by an operator on the 排气格栅 with nobody else in range it does not attack, as in the official game (community report of 2026-10-06, item 24)

2026-10-07 02:07:51 +08:00

feedback5-cb3-pins.test.js

feedback5 CB3: pin four community reports of 2026-10-06 this tree already answers — 铃兰's S3 trigger (item 2), the 突袭 玛恩纳 (16.1), 归溟幽灵鲨 devoured (16.4), 荒芜拉普兰德 / 叙拉古 (46)

2026-10-07 02:39:30 +08:00

feedback5-cb4-pins.test.js

feedback5 CB4: the moving leaders' routes checked on every boss field (item 29); 野鬃 pushing enemies off the map (38), 锏 pulling a weight-10 leader (39), 昆图斯's footprint as 3×3 (47) and the leaders' low attack rate (52) not reproduced — pinned, no change (community reports of 2026-10-06)

2026-10-07 02:40:50 +08:00

feedback5-chain-heal.test.js

feedback5 WE1: 链愈师 heal jumps within the 3×3 around the last one healed (x-4), a full-HP ally included (follow-up 27)

2026-10-06 16:00:20 +08:00

feedback5-community-pins.test.js

feedback5 WAA: DESIGN §25.14 — the four community reports of 2026-10-06 (two fixed, 忍冬's trigger and 枯朽之种's bounty already answered and pinned)

2026-10-06 22:05:47 +08:00

feedback5-countdown-summons.test.js

feedback5 CB2: the countdown summons (医疗探机, 诅咒娃娃, 海嗣; 自选 工程蓄水炮, 沙地兽, 打字机, 本能的召唤) take no damage — 活性源石 included —, are never healed, and their bar shows the life left (community report of 2026-10-06 「赫默的无人机、浊心斯卡蒂的海嗣等这种召唤物不会受到外来伤害…血量随时间减少」)

2026-10-07 02:27:33 +08:00

feedback5-counters-any-damage.test.js

feedback5 CB3: the "受到攻击时" counters answer every damage instance from an enemy, not its attacks only — 菲莱 S2 now blasts on 深溟巢涌者's pulse, so do 星熊 S2, 泡泡, 刺玫 S2, 余 S1, 玛恩纳 无动于衷, 年 S2 and 斩业星熊 (community report of 2026-10-06, item 30)

2026-10-07 02:30:07 +08:00

feedback5-enemy-swing.test.js

feedback5 WE1: enemy attacks strike at their damage frame; a short stun before it cuts the swing (GitHub #187 point 1, #170)

2026-10-06 15:21:24 +08:00

feedback5-frozen-reblock.test.js

feedback5 R19A: a stunned or frozen enemy is blocked by contact where it stands — an operator redeployed beside a frozen 隐匿 enemy blocks it and so reveals it (GitHub #232, @ccyl; the repro and the cause by @kukiC); golden: 52 scenarios move

2026-10-07 11:22:27 +08:00

feedback5-gravel-double-hit.test.js

feedback5 CB3: 砾's normal attack hits twice at 50 % 伤害倍率 after DEF, so a 频次 enemy loses 2 per attack (community report of 2026-10-06, item 16.5)

2026-10-07 02:12:37 +08:00

feedback5-groundbind.test.js

feedback5 O20: the 缚地 status (groundbind) — an enemy air unit counts as a ground unit and cannot move

2026-10-06 05:20:28 +08:00

feedback5-hpdamage-hook.test.js

feedback5 O5: damage.js gets the hpDamage hook — what passed the shields of a damage instance, before the HP loss; a handler may only lower it (a raise is ignored), a 流失 and a hit the shields absorb whole never reach it

2026-10-06 00:07:06 +08:00

feedback5-kjerag-cold.test.js

feedback5 CB1: a pair of colds on an enemy becomes the freeze, no cold is left — the 谢拉格 wind no longer freezes for good from 51 layers on (community reports 「谢拉格盟约冰冻时间没有随层数正确成长」 and 「…被在无法被任何干员攻击到的地方永控」, items 7 and 8)

2026-10-07 01:39:50 +08:00

feedback5-lemuen-s3-wait.test.js

feedback5 R19G: 蕾缪安's S3 no longer ends by itself when nothing is in her range — it keeps its bullets and locks and locks the next enemy that comes; the bombardment follows only when the bullets are spent (community report); golden: 3 scenarios move

2026-10-07 12:19:55 +08:00

feedback5-musha-heal.test.js

feedback5 WE1: 武者 trait heal on every damage instance, through 禁疗 (follow-up 1)

2026-10-06 15:32:30 +08:00

feedback5-mystic-store.test.js

feedback5 WE1: 秘术师 stores its energy at the attack check, not one interval later (GitHub #181)

2026-10-06 14:53:56 +08:00

feedback5-o8-engine.test.js

feedback5 O8: two additive engine rules for the 自选 kits — a 屏障 of one damage type and a profile's own target exclusion

2026-10-05 23:55:13 +08:00

feedback5-o16-engine.test.js

feedback5 O16: two additive engine rules for the 自选 kits of 机械师 and 凯尔希 — a 屏障 limited to a list of damage types, a healer that heals one ally through its 禁疗

2026-10-06 01:33:46 +08:00

feedback5-o23-engine.test.js

feedback5 O23: 自选 kits for 白铁, 娜斯提 (with their 支援装置) and 死芒 (with her 悲叹的仆役); ally targets for 白铁's 铁钳号·原型机

2026-10-06 05:28:54 +08:00

feedback5-raid-own-board.test.js

0.2.5-dev: 驭械术师 normal attacks keep the operator's own hit plus one drone hit, the drone multiplier and per-target map reset on redeploy, 洛洛 S2 keeps its lock until a kill or the skill ends (PRs #537 / #536 / #535 by @Xinglan233, ported); hit tiles use the official round-half-to-even with an edge inset (PR #520 by @Cloudnyco, ported); bots softly avoid living teammates' opening factions until they are eliminated (PR #518 by @IceCodeNew, ported with changes); golden: 43 more scenarios move, each named in docs/history/0.2.5.md

2026-10-12 14:45:01 +08:00

feedback5-raid-revive.test.js

feedback5: merge CB1 — community reports (bonds / strategies / items): a pair of colds on an enemy becomes the freeze and leaves no cold (the 谢拉格 wind no longer freezes for good), a 复活 (埃芒加德's 命结之秘, M3茧甲) is a knock-out and an immediate free redeploy where the operator lies (砾's and 瑕光's 被击倒时 / 部署时 layers fire), 佩佩's special refresh is paid, the result card hides the layers of 调和 / 协防干员 / 独行 / 绝技, the 炎佑 keeps off 外勤医疗's slot, 凯瑟琳's 定向投放 offers every shop item; 烛煌's 绝处重燃 before the hammer's lock and the 寻呼模块 / 简易通讯机 tier cap kept as they are (the owner's decisions of 2026-10-07); §25.15 → §25.21 (§25.15 is the shortcuts'); golden: vs feedback5 CB1's own rules — roster-005/009/010/012/014/015/016/018/026/029/031/033/036/040/042/046/047/049, bond-egirShip-high, bond-indomShip-low, bond-kjeragShip-high, bond-maniShip-high, bond-raidShip-low, bond-steadShip-high, bond-suntShip-low, boss-boss_2/3/4/6-pair, hidden-boss_8/9/10-pair, standin-07, matches solo-FUNNY-1, solo-NORMAL-1, solo-HARD-1, solo-ABYSS-1, coop3-ABYSS-5, coop4-FUNNY-6, coop4-ABYSS-7; vs CB1 what feedback5 changed since its base (CB2, CB3, CB4, CB5) plus roster-014

2026-10-07 03:54:35 +08:00

feedback5-ulpia-own-board.test.js

feedback6 R21B: 联防 plays on the round's battlefield again — its terrain, crates, water, devices and runes, as in 0.1.4 (the owner's decision of 2026-10-07, 「官服保留地形，改回去」; GitHub #244). 0.2.0's escaped-level map (8a436249, §25.6.4) is withdrawn: that map is the placeholder grid all 39 wave template levels carry, tile for tile; only the 11 stage levels have terrain. match/unitePhase.js _uniteOpts fields this.stageId in both combat modes, unite.js uniteStageId is removed, the client draws the 联防 on the board's stage again (screens/game.js: 0.1.4's ownStage / baseStage, no map switch, nothing to restore on leaving), the dev mock likewise. Kept from 0.2.0: own-board 突袭 / 乌尔比安 S3 landings (onOwnBoard, raidTile untouched), the bounty payee, the 联防 BGM, the SETTLE result box. data/stages.json keeps the two kind 'unite' records (no field uses them; engine tests use them as a plain two-halves road); tools/build-data.mjs comments only, an offline rebuild reproduces data/ byte for byte. The 联防 battle input is byte-identical to v0.1.4's in 20 cases (5 stages × 1 / 2 helpers × both combat modes); GitHub #41 item 3 (野鬃 pushing walkers back across the gate onto the other half) is back with v0.1.4's numbers, none off the map. Tests: test/match/feedback5-unite-map.test.js rewritten (6, all fail on the parent), test/e2e/coop.e2e.mjs checks the 联防 field's drawn stage; docs: DATA §12, META §4, PLAYING §5, design/geometry-time, I18N, the §25.6.4 pointer; golden: fields unite-1 (escaped_single → act1autochess_m02) and unite-2 (escaped_multi → act1autochess_m07), the stages of 0.1.4, and the co-op matches whose 联防 rounds come out differently on the round's map — coop2-NORMAL-3, coop2-HARD-4, coop3-ABYSS-5, coop4-ABYSS-7, coop2-NORMAL-8-serverrun, coop2-NORMAL-14-standins (now a victory)

2026-10-07 17:40:03 +08:00

feedback5-ulpia-revive-home.test.js

feedback5: 乌尔比安 knocked out while his S3 has moved him lies and redeploys on his deployment tile, not on the anchor tile — the owner's decision of 2026-10-07 on community report 28 「乌尔比安3技能期间死亡会留在3技能位移后的位置复活，应回到初始部署位复活」, a deliberate deviation from PRTS's "where it fell" for this 【移动】 only; golden: unchanged (the corpus moves him 23 times and knocks him out 6 times, never while moved)

2026-10-07 04:04:41 +08:00

feedback7-allytarget-cast.test.js

feedback7 R22-OSIM: 白铁's 铁钳号·原型机 counts like an enemy for a non-heal skill's automatic start; golden: unchanged

2026-10-08 19:36:35 +08:00

feedback7-raid-stealth.test.js

feedback7 R22-OSIM: 突袭 members jump to 隐匿 ground enemies too, and one in range counts for 「范围内没有敌人」 (the owner's decision, GitHub #316); golden: 5 match scenarios move

2026-10-08 21:21:44 +08:00

fidelity.test.js

feedback7 UIBACK: the top-bar capsule counts like the official — resolved = killed + leaked among the enemies the round scheduled (opening 0/3, a leak 1/3, a kill 2/3, the splitter 3/3); a split child, a summon and a boss summon are in neither number but still cost LP and break 完美作战 (PR #308 by @xcdoge, the capsule half of PR #157 — authored by them: the sim / protocol / HUD hunks applied as reviewed): inTotal / killedInTotal / leakedInTotal / resolved, b.snap / BattleResult / b.progress (optional field) / m.public progress carry resolved, killed may exceed total (validation bounds it by the content-spawn allowance); dropped the PR's stray design-doc line — it had replaced the S→C b.end bullet of docs/design/network.md, which is back; the 联防 live-counter change mark follows the enemy list now that a spawn no longer moves total (uniteLeftMark + test); golden: only total / players.*.total / snaps move, outcomes bit-identical — roster 23 (roster-002 003 007 008 009 010 011 012 013 015 017 018 020 023 024 025 026 027 037 045 047 048 049), bonds 2 (bond-maniShip-low, bond-soloShip-high), fields 11 (boss-boss_1 3 5 pair+solo, boss-boss_6-solo, hidden-boss_8 10 pair+solo), diy 11 (diy-036 040 056 062 068 073 083 091 095 114 115), matches 12 (solo-FUNNY-1 NORMAL-1 HARD-1 ABYSS-1 ABYSS-2, coop2-NORMAL-3 coop2-HARD-4 coop3-ABYSS-5 coop4-ABYSS-7 coop2-NORMAL-8-serverrun solo-HARD-9-boosted coop2-NORMAL-70-diy: the total column of 36 round rows), standins unchanged

2026-10-08 19:30:36 +08:00

funnel-body.test.js

0.2.5-dev: 耶拉 S2 (both forms) and 荒芜拉普兰德 S1 keep the operator's own hit next to the drone hit — only drone-only skills skip it (review fix of the funnel port); golden: more funnel scenarios move, attributed in docs/history/0.2.5.md

2026-10-12 15:50:41 +08:00

funnel-redeploy.test.js

0.2.5-dev: 驭械术师 normal attacks keep the operator's own hit plus one drone hit, the drone multiplier and per-target map reset on redeploy, 洛洛 S2 keeps its lock until a kill or the skill ends (PRs #537 / #536 / #535 by @Xinglan233, ported); hit tiles use the official round-half-to-even with an edge inset (PR #520 by @Cloudnyco, ported); bots softly avoid living teammates' opening factions until they are eliminated (PR #518 by @IceCodeNew, ported with changes); golden: 43 more scenarios move, each named in docs/history/0.2.5.md

2026-10-12 14:45:01 +08:00

fuzz.test.js

feedback7 UIBACK: the top-bar capsule counts like the official — resolved = killed + leaked among the enemies the round scheduled (opening 0/3, a leak 1/3, a kill 2/3, the splitter 3/3); a split child, a summon and a boss summon are in neither number but still cost LP and break 完美作战 (PR #308 by @xcdoge, the capsule half of PR #157 — authored by them: the sim / protocol / HUD hunks applied as reviewed): inTotal / killedInTotal / leakedInTotal / resolved, b.snap / BattleResult / b.progress (optional field) / m.public progress carry resolved, killed may exceed total (validation bounds it by the content-spawn allowance); dropped the PR's stray design-doc line — it had replaced the S→C b.end bullet of docs/design/network.md, which is back; the 联防 live-counter change mark follows the enemy list now that a spawn no longer moves total (uniteLeftMark + test); golden: only total / players.*.total / snaps move, outcomes bit-identical — roster 23 (roster-002 003 007 008 009 010 011 012 013 015 017 018 020 023 024 025 026 027 037 045 047 048 049), bonds 2 (bond-maniShip-low, bond-soloShip-high), fields 11 (boss-boss_1 3 5 pair+solo, boss-boss_6-solo, hidden-boss_8 10 pair+solo), diy 11 (diy-036 040 056 062 068 073 083 091 095 114 115), matches 12 (solo-FUNNY-1 NORMAL-1 HARD-1 ABYSS-1 ABYSS-2, coop2-NORMAL-3 coop2-HARD-4 coop3-ABYSS-5 coop4-ABYSS-7 coop2-NORMAL-8-serverrun solo-HARD-9-boosted coop2-NORMAL-70-diy: the total column of 36 round rows), standins unchanged

2026-10-08 19:30:36 +08:00

hit-tile-round.test.js

0.2.5-dev: 驭械术师 normal attacks keep the operator's own hit plus one drone hit, the drone multiplier and per-target map reset on redeploy, 洛洛 S2 keeps its lock until a kill or the skill ends (PRs #537 / #536 / #535 by @Xinglan233, ported); hit tiles use the official round-half-to-even with an edge inset (PR #520 by @Cloudnyco, ported); bots softly avoid living teammates' opening factions until they are eliminated (PR #518 by @IceCodeNew, ported with changes); golden: 43 more scenarios move, each named in docs/history/0.2.5.md

2026-10-12 14:45:01 +08:00

hud-capsule-resolved.test.js

feedback7 UIBACK: the top-bar capsule counts like the official — resolved = killed + leaked among the enemies the round scheduled (opening 0/3, a leak 1/3, a kill 2/3, the splitter 3/3); a split child, a summon and a boss summon are in neither number but still cost LP and break 完美作战 (PR #308 by @xcdoge, the capsule half of PR #157 — authored by them: the sim / protocol / HUD hunks applied as reviewed): inTotal / killedInTotal / leakedInTotal / resolved, b.snap / BattleResult / b.progress (optional field) / m.public progress carry resolved, killed may exceed total (validation bounds it by the content-spawn allowance); dropped the PR's stray design-doc line — it had replaced the S→C b.end bullet of docs/design/network.md, which is back; the 联防 live-counter change mark follows the enemy list now that a spawn no longer moves total (uniteLeftMark + test); golden: only total / players.*.total / snaps move, outcomes bit-identical — roster 23 (roster-002 003 007 008 009 010 011 012 013 015 017 018 020 023 024 025 026 027 037 045 047 048 049), bonds 2 (bond-maniShip-low, bond-soloShip-high), fields 11 (boss-boss_1 3 5 pair+solo, boss-boss_6-solo, hidden-boss_8 10 pair+solo), diy 11 (diy-036 040 056 062 068 073 083 091 095 114 115), matches 12 (solo-FUNNY-1 NORMAL-1 HARD-1 ABYSS-1 ABYSS-2, coop2-NORMAL-3 coop2-HARD-4 coop3-ABYSS-5 coop4-ABYSS-7 coop2-NORMAL-8-serverrun solo-HARD-9-boosted coop2-NORMAL-70-diy: the total column of 36 round rows), standins unchanged

2026-10-08 19:30:36 +08:00

hud-readouts.test.js

feedback7 R22-HUD: the HP bar shows the ammo skill's rounds as yellow cells (隐现 S2: 14), 伺夜's 狼影 left as diamonds and 斩业星熊's 我执 pool as a red bar — one snapshot change for PR #303 and #378 by @IceCodeNew and PR #383 by @Convey123; golden: unchanged

2026-10-08 19:22:08 +08:00

loadout-summons.test.js

Playtest #6 fixes: summons placed by hand, boss pool stall, bounties, 联防 leak counter, distinct offers, 机变 card text, phone buy tap, skill triggers, blocking, push force, 孤立, enemies, element bar

2026-10-01 19:08:23 +08:00

loadout.browser.test.js

Dev: puppeteer-core 25 (browser tests adapted)

2026-10-02 17:58:35 +08:00

loadout.test.js

feedback7 R22-LOADOUT: the battle takes each operator's potential and 练度 from its spec entry; golden: unchanged

2026-10-08 20:38:08 +08:00

morph-bonds.browser.test.js

0.2.5-dev: community round 30 and follow-ups

2026-10-12 14:07:08 +08:00

official-damage-numbers.test.js

0.2.5-dev: an optional 「官方伤害数字」 setting (default off) shows only the numbers the official game shows per the PRTS red-number rule (#412, the owner's decision of 2026-10-12); a small link to the project's GitHub repository on the title page (#543); golden: unchanged

2026-10-12 14:59:20 +08:00

palsy-trigger.test.js

feedback5 O3: 自选 operator kits for 逻各斯, 真言 and 维伊 (every skill, talent, trait and module at stages 1 / 3), the engine hook palsyTrigger

2026-10-06 00:04:18 +08:00

pathing-blockable.test.js

feedback7 R22-FIELDS: the path tie-break never wades into more 深水区 than the official route — on 战场#08(下半) the patrolling 铳 / 卢西恩 take the official upper road from (3,9) to (2,3) again (GitHub #375); golden: boss-boss_2-pair, coop2-ABYSS-10-boosted

2026-10-08 20:14:37 +08:00

pathing-crosscheck.test.js

feedback7 R22-FIELDS: the path tie-break never wades into more 深水区 than the official route — on 战场#08(下半) the patrolling 铳 / 卢西恩 take the official upper road from (3,9) to (2,3) again (GitHub #375); golden: boss-boss_2-pair, coop2-ABYSS-10-boosted

2026-10-08 20:14:37 +08:00

pathing-official.test.js

feedback7 R22-FIELDS: the path tie-break never wades into more 深水区 than the official route — on 战场#08(下半) the patrolling 铳 / 卢西恩 take the official upper road from (3,9) to (2,3) again (GitHub #375); golden: boss-boss_2-pair, coop2-ABYSS-10-boosted

2026-10-08 20:14:37 +08:00

pathing.test.js

feedback1d WO: official route wherever it crosses no more floor (D5 战场#04 diagonal)

2026-10-03 04:14:43 +08:00

perf.test.js

卫戍协议：盟约 web remake — initial import

2026-09-29 16:21:31 +08:00

playtest5_blocking.test.js

Playtest #6 fixes: summons placed by hand, boss pool stall, bounties, 联防 leak counter, distinct offers, 机变 card text, phone buy tap, skill triggers, blocking, push force, 孤立, enemies, element bar

2026-10-01 19:08:23 +08:00

playtest5_bossbody.test.js

Playtest #5 fixes: hovering enemies, target order, blocking hand-over, 联防 knocked-out, elements, huge-boss hit area, maps, gear icon, phone HUD, 至简 price

2026-09-30 19:12:36 +08:00

playtest5_elements.test.js

Playtest #5 fixes: hovering enemies, target order, blocking hand-over, 联防 knocked-out, elements, huge-boss hit area, maps, gear icon, phone HUD, 至简 price

2026-09-30 19:12:36 +08:00

playtest5-targeting.test.js

feedback7 R22-GHSIM: a 速射手 (优先攻击空中单位) blocking a ground enemy shoots a flyer in its range first; every other attacker keeps its blocked enemy first (GitHub #220, the owner's decision of 2026-10-08); golden: roster-026, roster-037 move

2026-10-08 20:29:45 +08:00

playtest6_bosspool.test.js

Playtest #6 follow-up (v2.5.1): leader HP review — 直接乘算, one debuff instance, 999 layer cap, 限伤; elite to the board; 假想敌：胄 kit

2026-10-02 04:19:28 +08:00

playtest6_elemgauge.test.js

Playtest #6 fixes: summons placed by hand, boss pool stall, bounties, 联防 leak counter, distinct offers, 机变 card text, phone buy tap, skill triggers, blocking, push force, 孤立, enemies, element bar

2026-10-01 19:08:23 +08:00

playtest6_limits.test.js

feedback5 WY: the Final Assault / Hidden Core leader pool is bloodPoint × the players alive when the fight starts (solo × 1), replacing 「保持固定血量」; 胄's drone share passes 限伤 (from PR #209 by @qingjingshenghuo; the owner's decision of 2026-10-06)

2026-10-06 21:41:49 +08:00

playtest6-combat.test.js

feedback6 R21P: every operator fights at full potential (潜能 6) — the owner's decision of 2026-10-07 (GitHub #252, PR #255; no host slider): tools/build-data.mjs OPERATOR_POTENTIAL adds every character_table potentialRanks attribute step (部署费用 / 再部署时间 / ATK / HP / DEF / RES / ASPD, all ADDITION, before rounding) to every chess (normal and elite), 补位 and 自选 unit form and picks talent / trait / module-talent candidates up to requiredPotentialRank 5 (「天赋效果增强」); a summon picks its talent candidates at its owner's potential (their requiredPotentialRank mirrors the owner's — PRTS 凯尔希: Mon3tr 3.5 s / 1400) and takes none of the owner's attribute steps; the 原型干员 stand-ins have no potential ranks (unchanged). 刺玫 413 / 17 → 435 / 15 as in the tournament video, 砾 18 → 16 s, 至简 80 → 70 s. data/chess.json (256 records), backups.json (the 71 owned picks, 8 summons), tokens.json (5 summons) and the data translations en / ja / ko / zh-TW rebuilt — only potential-driven fields (JSON-compared); the 干员调配 局内数值 and detail cards read the records. op-wang.js reads her stock cap 「最多拥有8（+1）枚」 through the potential step (the pattern fell back to 7: she sat at the cap from the start). Docs: DATA §2 / §18 / tokens, PLAYING (detail card), research 03, kits README, shared/diy.js / standIn.js, the 71 自选 kit headers and the kit comments that quoted potential-0 numbers. Tests: test/full-potential.test.js (new, fails on the parent); data.test.js re-derivation adds the potential steps (+ respawnTime); every test that hard-coded a potential-0 number carries the full-potential one (from the data and the test's own formula; no check loosened); golden: moves almost everywhere — roster 49/49, bonds 46/46, fields 22/22, matches 18/18, diy 133/138 (each battle parts on a stat: roster-001 / roster-020 at tick 1 on 角峰's HP (生命上限+250), bond-arcaneShip-high at 4.73 s and diy-001 at 3.70 s on an enemy's HP from a higher ATK; diy-110 / 111 also by 望's cap; the matches through those battles); unchanged: standins 10/10 and diy-135…138 (no potential ranks), diy-012 (twelve 夜莺 never hit, never redeployed)

2026-10-07 18:35:39 +08:00

playtest6-enemies.test.js

Playtest #6 fixes: summons placed by hand, boss pool stall, bounties, 联防 leak counter, distinct offers, 机变 card text, phone buy tap, skill triggers, blocking, push force, 孤立, enemies, element bar

2026-10-01 19:08:23 +08:00

playtest6-skills.test.js

Update expanded roster checks and register Clementia impact visuals

2026-10-09 17:46:07 +08:00

portal-appear.test.js

feedback7 R22-FIELDS: an enemy whose route ends on a corner portal stays hidden 3 s inside it before it comes out of [5,10], like the explicit crossings of the same tiles (review of the #337 commit); golden: 7 fields, 4 matches — solo-NORMAL-1's Hidden Core still fails

2026-10-08 21:05:18 +08:00

potential-battle.test.js

feedback7 R22-LOADOUT: the battle takes each operator's potential and 练度 from its spec entry; golden: unchanged

2026-10-08 20:38:08 +08:00

professions.test.js

feedback5 WL: 生命回复速度 is not 治疗 — the 吟游者 trait, 调香师, 瑕光 S2, 铃兰 S3 and 锡人 give an hpRegen buff, so 无法被友方治疗 / 禁疗 units get it and 治疗加成 does not scale it (GitHub #96 / #137)

2026-10-05 21:13:15 +08:00

raid-terrain.test.js

灵巧's knocked-out aura counts only knocked-out operators; hidden-layer bonds hide their layers in the UI only (#66)

2026-10-05 22:34:39 +08:00

robustness.test.js

feedback1 final: the boss-field speed test tolerates CI runners (1.0 ms with CI set, best of three runs; 0.5 ms locally)

2026-10-03 19:18:37 +08:00

simcheck.test.js

0.2.5-dev: community round 30 and follow-ups

2026-10-12 14:07:08 +08:00

skills.test.js

The skill-window test of #160 follows #109: a deploy-timed skill (缄默德克萨斯 S2) is a duration skill with its own start / end events and one charge per deployment

2026-10-05 23:18:53 +08:00

strongest-status.test.js

feedback7 R22-GHSIM: three or more overlapping 同名效果取最高 applications resume strongest first — the middle one no longer drops out when the strongest ends (GitHub #342, PR #348 by @siruimei07); golden: diy-059, diy-122, diy-123, diy-124, diy-133, diy-138 move

2026-10-08 18:48:26 +08:00

tick-tolerance.test.js

feedback7 R23-TICK: an attack interval or a time-SP cost of a whole number of frames takes exactly that many — 1 s is 30 frames, not 31 (PR #402 by @Cloudnyco, ported onto 0.2.2); golden: every family moves (roster, bonds, fields, matches, standins, diy — regenerated in the next commit)

2026-10-09 01:37:32 +08:00

trait_attack_speed.test.js

feedback7 R22-GHSIM: the module trait line 「攻击范围内存在N名及以上敌人时攻击速度+X」 (REA-Y: 圣约送葬人, 隐德来希) is applied by the engine instead of the two kits — on and off on the tick the count changes; a scan of every chess, 补位 and 自选 loadout shows no other operator gains it (PR #293 by @LimitlessHPPK); golden: unchanged

2026-10-08 18:56:12 +08:00

unbalance.test.js

feedback7 R22-UNBAL: 失衡 review fixes — a pulled 静态刚体 is held for the pull's window, 弧光锋卫 bleeds 400 every 0.066 s while the state lasts, no enemy skill / blink / scripted move during it, a route APPEAR or a 逐火 重生 ends it, every push row holds its exact frames; golden: 32 move, 8 in outcomes (below)

2026-10-08 22:42:26 +08:00