chore: delete 106 unreachable *_system.gd and their parity tests

These files were loaded only by their own test_*_parity.gd and never by
the runtime (transitively checked, incl. client_phase_integration_system).
Removes 76 manifest implementation/evidence references across 18
contracts; contract docs get a note that conclusions based on them are
void. Live systems kept: chest, consumable, item_attr, metin_socket,
whisper_chat.

Full headless suite: no new failures (7 failing + 5 headless-hanging
render tests are identical before the deletion).

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

audit/history.jsonl

@@ -484,3 +484,4 @@
 {"time": "2026-09-22T01:30:00Z", "event": "audit_round", "result": "PARTIAL", "ids": ["movement.keyboard.motion/canact-paralysis-faint-sleep-gate"], "files": ["audit/contracts/movement.keyboard-motion.md", "audit/manifest.json", "audit/history.jsonl"], "tests": ["40250 reference source reachability audit (exhaustive grep -arn across UserInterface/ + GameLib/ for SetParalysis/Paralysis, AFFECT_SLEEP/SetSleep, SetFaint/FaintTest, Stun/RecvStunPacket)"], "reason": "Investigated the movement.keyboard-motion.md Remaining item \"CActorInstance::CanAct() Paralysis/Faint/Sleep movement gates not wired into net_play.gd\" (flagged unfixed by the prior round). Findings: (1) IsParalysis()/m_isParalysis has exactly one setter in the whole reachable 40250 source, CInstanceBase::__Shaman_SetParalysis (InstanceBaseEffect.cpp:821-823), which itself has zero callers anywhere in the tree -- Paralysis is dead code, never true in live gameplay, so not wiring it is correct parity, not a gap. (2) IsSleep()/m_isSleep: the AFFECT_SLEEP case in SetAffect()s switch is commented out (dead); the only live setter is the AFFECT_STUN case (m_GraphicThingInstance.SetSleep(isVisible), ~InstanceBaseEffect.cpp:932-933) -- IsSleep() is functionally just an alias for AFFECT_STUN in practice, which the current clients Entity.stunned field already represents (entity_store.h:119) and net_play.gd::_can_process_network_state() already gates movement on (dead/stunned/knock_down) -- already covered, not a gap. (3) IsFaint()/m_isFaint: only setter is chrFaintTest() (PythonCharacterModule.cpp:1087-1105), a Python-bound debug helper (\"FaintTest\", registered alongside chrtestRestoreRenderMode) operating on GetSelectedInstancePtr(); the available 40250 checkout has no root/ Python UI/quest scripts to prove or disprove live-gameplay reachability -- evidence-blocked, left unresolved per SKILL.md do-not-audit-unreachable-code-as-product-behavior rule, no fix attempted. Byproduct finding while tracing IsStun(): CActorInstance::Stun() is only reachable via RecvStunPacket() (PythonNetworkStreamPhaseGame.cpp:1556-1578), which explicitly routes GC_STUN to Die() for the main character and to Stun() only for remote actors; verified this does NOT create a movement.keyboard.motion discrepancy because CanAct() blocks movement identically on IsDead() and IsStun(), and net_play.gd::_can_process_network_state() already blocks movement identically on dead and stunned -- the movement-blocking outcome is already correct parity regardless of which reference path fires. The Die()-vs-Stun() distinction only affects death-system semantics (m_isRealDead, EXP drop, corpse/respawn flow), which is combat.death.exp_drop/combat.affect_status territory and is already tracked there (combat.affect-status.md Remaining #2, equivalence rows 63/67: native GC_STUN only sets stunned=true with no dedicated clear path) -- not re-logged as a new finding to avoid duplicate tracking. No code changed this round; removed the resolved Remaining bullet from manifest.json and movement.keyboard-motion.md, replaced with investigation evidence and citations. Contract stays PARTIAL (root-motion/resource-event-source/sync-hit-death-gate-stacking/real-model-bone-flip items remain open)."}
 {"time": "2026-09-22T03:15:00Z", "event": "implementation_fix_round", "result": "PARTIAL", "ids": ["movement.keyboard.motion/server-speed-scale-clamp"], "files": ["project/player_controller.gd", "project/keyboard_motion_timeline_test.gd", "audit/contracts/movement.keyboard-motion.md", "audit/manifest.json", "audit/history.jsonl", "audit/remediation-roadmap.md", "audit/reports/coverage.md"], "tests": ["40250 reference source inspection (PASS: InstanceBaseMovement.cpp CInstanceBase::SetMoveSpeed(UINT uMovSpd) only guards uMovSpd>1100 -> 0; otherwise applies uMovSpd/100.0f unclamped)", "extension/src/net/entity_store.cpp EntityStore::motion_move_speed re-read (PASS: already-fixed remote-entity path mirrors the same >1100 -> 0 / mount_vnum!=0 -> 0 guards, no [0.25,3.0]-style band)", "pre-fix player_controller.gd::set_server_speed (FAIL by construction: clampf(moving_speed/100.0, 0.25, 3.0) silently floors/ceils haste or slow stacks outside the reference's 25..300 movSpd band)", "post-fix godot --headless --path project --script keyboard_motion_timeline_test.gd (PASS, incl. new set_server_speed(10/500/1200/100) assertions)", "godot --headless --path project --script test_wasd_steering_parity.gd (PASS)", "godot --headless --path project --script test_no_auto_move_regression.gd (PASS)", "godot --headless --path project --script game_camera_test.gd (PASS)", "godot --headless --path project --script movement_parity_test.gd (PASS)", "godot --headless --path project --script player_move_test.gd (PASS)", "godot --headless --path project --script mouse_controller_test.gd (PASS)", "godot --headless --path project --script test_alignment_parity.gd (PASS)", "godot --headless --path project --script netplay_test.gd (PASS; unrelated mock set_server_speed in netplay_test.gd has no clamp, unaffected)", "git diff --check (PASS)", "audit_ledger.py refresh --write/report --write/validate (PASS)"], "reason": "在同一合同下追查『根运动 vs 连续速度』假设时发现这是假线索——entity_store.cpp::advance_walk_by_motion 与本地 _process() 的位移积分在种类上同构（都是恒速直线插值）；真正的差异是 set_server_speed() 把 moving_speed/100.0 塞进无参考依据的 clampf(0.25, 3.0)，而参考实现 CInstanceBase::SetMoveSpeed 和本项目已修复的 EntityStore::motion_move_speed 都只有 moving_speed>1100 -> 0 这一个门。移除该私有夹钳，改为 0.0 if moving_speed > 1100 else moving_speed/100.0，两处调用方（net_world.gd、net_play.gd）均不额外夹钳，改动范围限于 player_controller.gd。新增回归断言覆盖旧边界内外和 1100 冻结门。副产品发现：CInstanceBase::SHORSE::SetMoveSpeed 是坐骑上玩家的独立速度设定，只在 IsMounting() 时生效，与远端 mount_vnum!=0 -> 0 的处理方式不同——本轮未下定论，留作独立候选。未修改根运动/AccumulationMovement 适配层、资源与网络事件完整顺序、同步/受击门叠加或真实模型测试，合同保持 PARTIAL。"}
 {"time": "2026-09-22T06:00:00Z", "event": "port_round", "unit": "UserInterface/PythonPlayerEventHandler.cpp", "ported": ["CPythonPlayerEventHandler::OnMove", "CPythonPlayerEventHandler::OnMoving", "CPythonPlayerEventHandler::OnStop"], "deleted": ["net_play.gd _tick_on_waiting idle FUNC_WAIT resend"], "divergent": ["CPythonPlayerEventHandler::OnWaiting (walking-but-stalled path not ported)"], "needs_live": [], "tests": ["project/netplay_test.gd PASS", "project/keyboard_motion_timeline_test.gd PASS", "project/movement_parity_test.gd PASS", "project/player_move_test.gd PASS"]}
+{"time":"2026-09-22T12:00:00Z","event":"dead_code_removal","deleted_system_files":106,"deleted_tests":106,"lines_removed":36221,"manifest_refs_removed":76,"contracts_touched":18,"reason":"project/*_system.gd reachable only from their own test_*_parity.gd; not implementations of 40250 behavior (SKILL.md: unreachable code is not evidence)","tests":["full headless Godot suite: 192 PASS, 7 FAIL and 5 render tests hanging in headless, all identical on the pre-deletion tree"]}