Root cause (pinned by the worktree investigation + verified against
source): CMotionInterp's dispatch tails strip link animations for
DETACHED objects only (retail `if (physics_obj->cell == 0)
RemoveLinkAnimations`, raw @305627, three ported sites:
DoInterpretedMotion / StopInterpretedMotion / StopCompletely). The R3
port proxied retail's cell pointer with CellPosition.ObjCellId == 0 -
but that #145 field is seeded ONLY by SnapToCell, whose single
production caller is the LOCAL PLAYER's SetPosition. Every REMOTE body
therefore read "detached" forever, and every dispatch stripped the
transition link the motion table had just appended: a used door's
swing link died the same tick (pose snapped closed<->open - the live
symptom), and remote walk<->run link poses have been silently eaten
since the guard landed (masked: locomotion cycles resemble each other;
a door's two poses don't). The dispatcher itself (GetObjectSequence
link path) was verified retail-verbatim end-to-end - not the bug.
Fix: PhysicsBody.InWorld - an explicit "placed in the world" flag
carrying exactly the guard's retail meaning (register row TS-40,
replacing the previously UNREGISTERED ObjCellId proxy). Set by
SnapToCell (retail: enter_world/set_cell assigns the cell) and by
RemoteMotion construction (a RemoteMotion exists only for a world
entity). All three guard sites now test !InWorld. Default false =
retail's pre-enter_world detached state, preserving the guard for
genuinely unplaced bodies (bare-harness interps unchanged).
Tests: InWorldLinkGuardTests pins the polarity at all guard classes
(in-world dispatch keeps links; detached strips; StopCompletely same).
Full suite green: 3,961.
If the door still snaps after this, the open question is DATA (does
the door's table author an On<->Off link at all) - next step per the
investigation: dump the real table's Links/StyleDefaults via a
DoorSetupGfxObjInspectionTests extension.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Live bug (2026-07-03, user door test): a server MoveTo armed
(movingTo=True) but the local body never moved; the retail observer saw
ACE walk the server-side player to the door; the next local input
reasserted the stale position (rubber-band). Log evidence: the player's
pending_motions queue was non-empty on 92/94 MotionDone pops (remotes:
0/40) - MoveToManager's retail wait-for-anims gate
(BeginTurnToHeading's MotionsPending check) never opened.
ROOT CAUSE (found via a production-faithful repro test - the original
suite force-drained the queue, masking it): the R3 port MISIDENTIFIED
retail's StopCompletely_Internal body. It is NOT a physics velocity
zero - CPhysicsObj::StopCompletely_Internal (0x0050ead0) tailcalls
CPartArray::StopCompletelyInternal (0x00518890) which is
MotionTableManager::PerformMovement(type 5): an ANIMATION-side full
stop whose UNCONDITIONALLY-queued pending_animations entry is the
completable partner of the A9 pending_motions node
CMotionInterp::StopCompletely enqueues (raw @00527e90 between the
state writes and add_to_queue). Without the dispatch, every
StopCompletely (login/teleport SetPosition + every MoveToManager
PerformMovement head) left an orphan node with no completion - the
queue never drained at idle. (The R3 code comment claimed a register
row for the stand-in; none was ever added - porting the real mechanism
closes that violation without a row.)
Fix set (all retail-anchored):
- IInterpretedMotionSink gains StopCompletely() (default true = the
null-sink posture); MotionTableDispatchSink routes it to the R2
manager's already-ported type-5 op (StopObjectCompletely +
Ready-sentinel entry) + TurnStopped for the ObservedOmega seam.
- MotionInterpreter.StopCompletely calls DefaultSink?.StopCompletely()
at the exact retail slot (after the state writes, before
add_to_queue); the immediate set_velocity(Zero) is kept as the
documented physics-effect stand-in.
- PlayerMovementController ticks CheckForCompletedMotions after
UseTime - retail's per-tick CPartArray::HandleMovement slot
(0x00517d60 = MotionTableManager::UseTime = CheckForCompletedMotions
tailcall, called every tick from UpdateObjectInternal @005159a4).
NOTE: a first-cut per-tick apply_current_movement pump was tried and
REVERTED - retail's apply callers are all event-driven (hold-key,
HitGround/LeaveGround, ReportExhaustion), not per-tick; the remote
block's per-tick apply is a pre-existing adaptation outside this
fix's scope.
- The P1 autonomous-store family, completing the pin's port shape
(the gate landed in V5; the STORE side was missing):
(a) the unpack store (00509730: last_move_was_autonomous = wire
byte) in the GameWindow player branch - local player only for now
(remote interps have no WeenieObj, which A3 reads as IsThePlayer;
storing a remote's autonomous byte would mis-route their per-tick
apply onto the raw branch);
(b) input edges store =1 at the controller edge block - retail's
CPhysicsObj::DoMotion @00510030 / StopMotion @005100e0 /
TakeControlFromServer @006b32f4 input-boundary stores;
(c) section 2's per-frame velocity write now PRESERVES the flag
(V5's first cut stamped it per frame, the wrong altitude - retail
stores it at event boundaries only);
(d) InstallSpeculativeTurnToTarget stores false (models the wire
mt-6's unpack store; part of the AP-23 adaptation) so the
event-driven applies (e.g. HitGround mid-moveto) route interpreted
and cannot clobber the manager's dispatched motions with idle raw
state.
Tests: new ServerMoveToPosition_ProductionCompletionFeed_WalksToArrival
drives the controller under the PRODUCTION completion feed
(sequencer.Advance -> MotionDoneTarget) instead of force-draining -
this is the test that reproduced the live wedge (queue stuck at the
two A9 Ready orphans) and now proves the drain. Existing cutover tests
mirror the wire-path autonomous store before PerformMovement. Full
suite green: 3,957.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
MovementParameters gains the verbatim selection family:
GetCommand 0x0052aa00 (the walk-vs-run cascade INCLUDING the CanCharge
0x10 fast-path ACE dropped - retail's default can_charge=false + the
fast-path present, the A13+A15 canceling-pair trap avoided; inclusive
threshold edge per the raw), TowardsAndAway, GetDesiredHeading per the
live Ghidra decompile (fwd-towards 0 / fwd-away 180 / back-towards 180
/ back-away 0), FromWire/FromWireTurnTo (UnPackNet semantics, all 18 A4
masks round-tripped).
New MoveToMath: HeadingDiff per the live Ghidra decompile of 0x00528fb0
(the 360-diff NOT-TurnRight mirror + F_EPSILON 0.000199999995f - the
BN "arg unused" artifact corrected), HeadingGreater (the visible
TurnRight idiom), PositionHeading/Get/SetHeading reusing the codebase's
single yaw-heading convention (P5), CylinderDistance (PDB arg order;
planar-minus-radii shape documented as the interpretation - the raw's
x87 body is garbled; seam noted).
MovementType gains Invalid + retail 6/7/8/9; MovementStruct widened
(ObjectId/TopLevelId/Pos/Radius/Height/Params, additive); WeenieError
+= 0x0B/0x36/0x37/0x38/0x3D with retail-meaning doc comments.
148 new conformance tests. Full suite: 3,860 passed.
Implemented by a dedicated agent against the V0-pinned spec; scope +
suite independently verified.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
User live report (the R3 visual pass): pressing W stopped the local
player instantly (retail observers saw run + rubber-band back — our AP
never moved); S/strafe appeared to work. The edge tracer + a
harness-with-echo repro converged on the funnel:
ApplyInterpretedMovement live-read InterpretedState.ForwardCommand
AFTER the style dispatch. The style dispatch SUCCEEDS against the real
MotionTableDispatchSink (GetObjectSequence Branch 1 style==target →
success — verified in the raw @298636), which gates the
ModifyInterpretedState write → InterpretedMotionState::ApplyMotion's
style branch resets forward_command to Ready UNCONDITIONALLY (raw
0051ea6c — our port is verbatim). Retail SELF-HEALS: its compiled
apply pass reads the axis fields into registers BEFORE the style call,
so the fwd dispatch re-applies the pre-reset command and the field
recovers within the pass — proven by our own 183-case live-retail
observer trace (the fwd dispatch carries the wire's RunForward after
the style dispatch on the same UM). The BN pseudo-C's apparent
post-style field reads at 0x528687 are decompiler rendering of hoisted
registers — the same artifact class as the A1 polarity inversion.
Under the live-field read, every apply pass (the ~10Hz ACE UM echo via
ApplyServerRunRate included) dispatched Ready and left the field
permanently Ready; the controller's per-frame get_state_velocity then
zeroed the body. The 183-case suite could not catch it: its
RecordingSink's return value doesn't mirror the real sink's
style-success, so the resetting state-write never ran under test.
Fix: entry-cache fwd/sidestep/turn commands+speeds (and the my_run_rate
read) before the style dispatch — restoring the S2a funnel's original
semantics through the W5 merge.
3 new regression tests bind the REAL sink over a real sequencer
(the masked condition): a full apply pass self-heals forward; 2 seconds
of held-W with echo cadence covers meters not centimeters; Shift
walk↔run toggling survives mid-transition echoes. Full suite: 3,734
passed. Temp diagnostics removed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The local player now drives the SAME pipeline remotes use. The
controller's per-frame RawMotionState rebuild (level-triggered D6.2) is
replaced by retail's CommandInterpreter altitude: key EDGES fire
DoMotion/StopMotion (0x00528d20/0x00528530) and the Shift edge fires
set_hold_run (0x00528b70, caller-0x006b33ca shape) — the interpreter's
own RawState mutates via ApplyMotion/RemoveMotion and every dispatch
routes the funnel + DefaultSink into the sequencer. GameWindow's
UpdatePlayerAnimation (169 lines) is DELETED with both
LocalAnimationCommand/LocalAnimationSpeed MovementResult fields and all
three synthesis sites (K-fix5 pacing, Walk→Run cycle selection, the
auto-walk overrides): run pacing now comes from apply_run_to_command's
my_run_rate promotion — the same source remotes use; airborne-Falling
falls out of contact_allows_move (the funnel's second dispatch).
Auto-walk's #69 turn-phase legs re-expressed as DoMotion(Turn*) edges;
teleport idle = StopCompletely (retail's full stop) + an edge-tracker
reset so held keys walk straight out of a portal.
Map discovery R1 fixed: ChargeJump() now fires at charge START (the
0x0056afac input-boundary site) — StandingLongJump had NEVER armed in
production despite the W3 port.
TWO root-cause fixes the cutover surfaced (each would have been
invisible under the old synthesis layer):
- CurrentHoldKey was a shadow FIELD synced only on the raw dispatch
branch; set_hold_run writes RawState.CurrentHoldKey, so an edge-driven
Shift toggle followed by DoMotion read a stale None and lost the run
promotion. Now an alias of RawState.CurrentHoldKey (retail's
adjust_motion reads raw_state.current_holdkey directly).
- The controller's grounded velocity + jump-launch writes used
set_local_velocity's default autonomous:false, silently clearing
LastMoveWasAutonomous and flipping the A3 dual dispatch to the
interpreted branch for the local player. Both marked autonomous.
EXPECTED-DIFFS (map §6, for the R3 visual pass): #45's 1.248x sidestep
factor + ACDREAM_ANIM_SPEED_SCALE died with UpdatePlayerAnimation —
local sidestep pacing now matches how remotes have always played;
auto-walk-at-run plays walk-pace legs until R4; ApplyServerRunRate's
apply_current_movement goes live (Branch-2 fast re-speed absorbs the
echo — retail's own mid-run speed-change mechanism).
Wire: outbound values stay input-derived (the L.2b-verified byte
stream untouched; RawState-sourcing needs the map §2b cdb TODO and R7
owns outbound). MotionStateChanged = the old comparison minus the dead
localAnimCmd leg, OR any edge fired.
Full suite green: 3,731 passed. Live smoke: in-world, user-driven
input through the new path (walk/turn/strafe), 12k entities, zero
exceptions.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The two parallel DoInterpretedMotion implementations MERGE into one
verbatim pair (0x00528360/@305639): contact gate, StandingLongJump
state-only branch, Dead → RemoveLinkAnimations, the jump_error_code
double-check WITH the DisableJumpDuringLink 0x20000 leg (W2's TODO(W5)
resolved — MovementParameters now flows), ModifyInterpretedState
gating, CurCell-null tail, AddToQueue with the real ContextId;
StopInterpretedMotion's post-stop Ready node + raw-mirror RemoveMotion.
Legacy overloads + ApplyMotionToInterpretedState DELETED. The funnel's
public surface unchanged (183-case suite compiles + passes as-is).
DoMotion 0x00528d20 verbatim: cancel_moveto interrupt; SetHoldKey(key,
cancelMoveToBit) BEFORE adjust_motion; params re-default for the
interpreted call; combat-stance gates on the ORIGINAL id (Crouch/Sit/
Sleep → 0x3f/0x40/0x41; & 0x2000000 → 0x42 outside NonCombat); action
depth cap ≥6 → 0x45; raw mirror ORIGINAL id. StopMotion 0x00528530
mirror shape. StopCompletely 0x00527e40 with the A9 snapshot quirk
(motion_allows_jump on the OLD forward command, stashed in the queued
node) + the J9 fix (sidestep/turn SPEEDS untouched — the 1.0 resets
were a divergence). PerformMovement flushes zero-tick completions after
every dispatched op via the new CheckForCompletedMotions seam
(0x0050fe30) — bound App-side per entity (remotes via
EnsureRemoteMotionBindings, player at the sequencer bind site).
PORT DISCOVERY (not in the W0 pins): retail's DoInterpretedMotion
RESULT gates both add_to_queue AND the state write — a void sink let
the style dispatch's apply-only failure clobber ForwardCommand before
the forward axis read it, regressing 74/183 live-trace cases.
IInterpretedMotionSink.ApplyMotion/StopMotion now return bool
(documented on the interface with the raw anchors); one funnel
assertion corrected with raw-line citations (airborne ForwardCommand
ends at Falling under the fully-wired verbatim algorithm).
62 new gate-table tests. Full suite: 3,729 passed.
Implemented by a dedicated agent against the W0-pinned spec; seam
bound + suite independently re-verified by the orchestrator.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Core (dedicated agent, independently reviewed): HitGround 0x00528ac0 /
LeaveGround 0x00528b00 verbatim (creature+gravity gates, the
RemoveLinkAnimations seam — K-fix18's retail mechanism — velocity via
GetLeaveGroundVelocity with the autonomous flag, jump-state resets,
apply_current_movement re-sync); enter_default_state 0x00528c80 per A8
(fresh states, InitializeMotionTables seam, sentinel APPENDED without
draining pending_motions — pinned, Initted=1, LeaveGround tail);
Initted gates; the A3 IsThePlayer dual dispatch in
apply_current_movement / ReportExhaustion / SetWeenieObject /
SetPhysicsObject (a remote player routes INTERPRETED — the
ACE-divergence pin); set_hold_run 0x00528b70 + SetHoldKey 0x00528bb0
(XOR guard, None-only-from-Run); adjust_motion creature guard wired
(TS-34 retired); PhysicsBody.LastMoveWasAutonomous +
set_local_velocity(autonomous). Port discovery: retail's
apply_raw_movement 0x005287e0 / apply_interpreted_movement 0x00528600
ARE the already-shipped D6.2a/funnel functions — the dual dispatch
composes them instead of duplicating.
App cutover (orchestrator): the skipTransitionLink flag + both K-fix18
call sites DELETED (AP-74 retired). MotionInterpreter.DefaultSink routes
apply_current_movement's interpreted branch through the REAL funnel
dispatch when a sink is bound — so a remote's LeaveGround engages
Falling via the contact-gated funnel, replacing the forced SetCycle
(J19); the per-remote MotionTableDispatchSink is now PERSISTENT
(EnsureRemoteMotionBindings: DefaultSink + RemoveLinkAnimations +
InitializeMotionTables seams, idempotent from both the UM and
VectorUpdate paths; wire velocity re-applied after LeaveGround so it
stays authoritative). Player: seams bound to the player sequencer; the
controller's grounded→airborne EDGE now fires LeaveGround (jump()
clears OnWalkable and the same frame's transition detection fires it —
retail's order; walk-off-a-ledge gets the momentum fallback + link
strip it never had); the manual jump-block LeaveGround deleted;
LastMoveWasAutonomous set at the controller chokepoint (W6 refines).
Trace S8 re-expressed as the retail mechanism (Falling dispatch +
RemoveAllLinkAnimations = same final state the flag produced). 43 new
lifecycle tests. Registers: TS-34 + AP-74 retired; TS-38, AP-77, AP-78
added. Full suite: 3,665 passed. Live smoke: in-world clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
JumpChargeIsAllowed 0x00527a50 (CanJump→0x49 before posture; Fallen +
Crouch..Sleeping→0x48), ChargeJump 0x005281c0 — now the ONLY place
standing_longjump arms (grounded Contact+OnWalkable + forward Ready +
no sidestep/turn, raw @305453-305466); the S2a-flagged side effect
inside contact_allows_move is DELETED (J6) with regression pins that
contact checks never touch the flag (the funnel's StandingLongJump
branch is local-player scope; the controller wires ChargeJump in W4/W6
— remotes never charge client-side, no interim regression).
jump_is_allowed 0x005282b0 full chain replaces the 15-line
approximation: IsFullyConstrained→0x47 (new PhysicsBody stub, TS-35)
BEFORE the pending-head peek (A2 ordering, pinned); head peek fires
whenever the queue is non-empty (no ACE Count>1 gate; nonzero
jump_error_code short-circuits the whole chain — test proves
JumpStaminaCost is never consulted); retail entry shape verbatim
(non-creature-weenie and gravity-off skip the ground gate; null
physics obj → 0x24 NOT 8 per A10); charge → MotionAllowsJump(fwd)
double-check → JumpStaminaCost gate (new IWeenieObject member,
always-affordable PlayerWeenie stub — TS-5 extended).
GetJumpVZ/GetLeaveGroundVelocity: epsilon corrected to retail's
0.000199999995f (was 0.001 — regression-pinned at extent 0.0005);
A5 shape (clamp 1.0, weenie-null 10.0f); A6 momentum fallback fires
only when ALL THREE components are sub-epsilon and overwrites all
three with global→local(m_velocityVector).
Jump 0x00528780: InterruptCurrentMovement no-op seam (TS-36, →R4
cancel_moveto), fires unconditionally; standing_longjump cleared ONLY
on failure (success keeps it — pinned). IWeenieObject.IsThePlayer()
(PlayerWeenie true) lands for W4's A3 dispatch.
51 new tests + 1 re-pin; superseded pre-W3 jump tests removed. Full
suite: 3,623 passed. Registers: TS-5 extended, TS-35/36/37 added.
Implemented by a dedicated agent against the W0-pinned spec; arming
gates + entry shape verified against the quoted raw text.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
MotionInterpreter implements IMotionDoneSink: pending_motions
(LinkedList<MotionNode>, AD-34 wording) + add_to_queue 0x00527b80 +
motions_pending 0x00527fe0 + MotionDone 0x00527ec0 verbatim (action-class
head → UnstickFromObject no-op seam (→R5 StickyManager) + BOTH states'
action-FIFO pops; head popped UNCONDITIONALLY — A7: never match-by-id,
params unread in this build) + HandleExitWorld 0x00527f30 (the raw
decompile's null-physics-obj branch would infinite-loop if translated
literally — adjudicated as retail dead code, documented on the method) +
is_standing_still 0x00527fa0 + MotionAllowsJump 0x005279e0 with the
A1-pinned literal blocklist (28-case boundary table test: Fallen blocks,
Falling passes — ACE's transposition NOT copied).
Queue producers wired into the funnel per the decomp anchors
(re-verified from the raw text): DispatchInterpretedMotion's success
path computes jump_error_code via retail's double-check shape
(motion_allows_jump(motion), then forward_command when non-action;
the DisableJumpDuringLink params leg is TODO(W5) — no MovementParameters
threaded yet) + add_to_queue; apply_interpreted_movement's turn-stop
re-queues the Ready node (@305766-305785 — the plan's producer #2 and #3
are the SAME site; StopInterpretedMotion's internal add_to_queue @305657
is W5 scope). ContextId=0 with TODO(W5): retail's own compiled code
reads an UNINITIALIZED local for it at this call site.
GameWindow: MotionDoneTarget now binds the entity's REAL consumer —
player via PlayerMovementController.Motion (new internal property),
remotes via RemoteMotion.Motion — resolved AT FIRE TIME (a remote's
RemoteMotion can be created after its first anim tick; an eager capture
would drop completions forever). Despawn runs BOTH layers' exit-world
drains (manager then interp) per §4. AD-36 narrowed (consumed for
creature-class; doors/statics recorder-only until R5).
51 new conformance cases incl. the end-to-end chain test:
MotionTableManager.AnimationDone → sink → interp pending head pops in
step. 183-case observer-trace + funnel suites green unchanged (queue
side effects additive). Full suite: 3,582 passed.
Implemented by a dedicated agent (MotionInterpreter side) against the
W0-pinned spec; producer placements + MotionAllowsJump branch algebra
independently reviewed against the raw decomp; GameWindow wiring by the
orchestrator.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
InterpretedMotionState + the unified RawMotionState (LegacyRawMotionState
folded away) gain retail's action FIFO + ApplyMotion/RemoveMotion, ported
verbatim from the raw named decomp (0x0051e790/0x0051eb60 region, pc
293252-293703) including two genuine retail quirks pinned by tests and
independently re-verified against the raw text before commit:
- RawMotionState::ApplyMotion's RunForward (0x44000007) dead branch — a
cycle-class apply of literal RunForward writes NOTHING (retail encodes
run as WalkForward + HoldKey_Run on the raw state; same family as the
D6 apply_run_to_command early-return).
- InterpretedMotionState::RemoveMotion's exact-match-only TurnRight/
SideStepRight handling (left variants fall through to the style/
forward branches).
MovementParameters verbatim (A4 pin): 18 named flags per the absolute
mask table, ctor = 0x1EE0F expansion + distance_to_object 0.6 /
fail_distance FLT_MAX / speed 1 / walk_run_threshhold 15 / hold_key
Invalid — with the two ACE-divergence traps ported RETAIL-side
(can_charge FALSE where ACE defaults true; threshold 15.0 not 1.0).
MotionNode {ContextId, Motion, JumpErrorCode} (acclient.h:53293) — the
pending_motions node W2 consumes.
WeenieError renumbered to retail's numeric semantics (A10 table):
NotGrounded=0x24, GeneralMovementFailure 0x24→0x47, new 0x3f/0x40/0x41
combat-stance rejects + 0x42 chat-emote + 0x45 action-depth; the
airborne jump/action returns corrected 0x48→0x24 per the A10 sweep
(incl. jump_is_allowed's null-physics-obj 0x24-not-8 divergence from
ACE). Codes are local-only — wire untouched.
Outbound packer proof: RawMotionStatePacker unmodified; all 6
golden-byte RawMotionStatePackTests pass unmodified. TS-24 register row
updated (capability landed; outbound wiring still open).
Implemented by a dedicated agent against the W0-pinned spec; quirks,
scope, and suite independently verified. Full suite green: 3,531
(374+425+713+2015+4 pre-existing skips).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Port the inbound funnel verbatim into MotionInterpreter:
- MoveToInterpretedState (0x005289c0): raw-state style adopt, FLAT
copy_movement_from overwrite, apply, then action replay under the
15-bit server_action_stamp wraparound gate (local player skips its
own autonomous echoes).
- ApplyInterpretedMovement (0x00528600): my_run_rate cache from
RunForward speed, then retail dispatch order style -> forward-or-
Falling -> sidestep(-stop) -> turn(-stop), turn early-return.
- DispatchInterpretedMotion (0x00528360): contact-gated sink dispatch
(IInterpretedMotionSink = the GetObjectSequence backend the App
implements); blocked non-action motions take the apply-only path —
retail's real mechanism behind K-fix17's 'airborne remotes keep
their cycle' empirical guard.
- contact_allows_move REWRITTEN VERBATIM from the real 0x00528240
(pseudo-C 305471): Falling/0x40000011 + turns always allowed,
non-creature weenies + no-gravity bypass, else Contact+OnWalkable.
The previous body conflated jump_charge_is_allowed (0x00527a50)
posture checks — the 2026-06-04 deep-dive divergence, now retired.
Six tests that pinned the misattributed behavior corrected
(grounded posture does NOT block motion; airborne accepts
Falling/turns only).
- IWeenieObject.IsCreature (default true) for the gate's non-creature
bypass.
Conformance harness (the user-requested 'prove it equals retail'
apparatus, layer 1): RetailObserverTraceConformanceTests parses the
LIVE cdb trace of a retail observer (Fixtures/l2g-observer-trace.log,
captured via tools/cdb/l2g-observer.cdb) into 183 golden cases —
each [MTIS] input state replayed through our funnel must produce
retail's exact [DIM] dispatch sequence. 183/183 conformant, including
the airborne jump case (replayed contact-free; pre-gate vs post-gate
accounting + HitGround second-pass truncation documented in-test).
13 synthetic funnel tests cover the branches the trace missed
(actions, stamps, longjump, sidestep).
GameWindow integration (S2b) follows; funnel not yet wired.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Ports retail's raw->interpreted motion normalization and routes the local
player's velocity through it, so backward/strafe come from the retail source
instead of hand-mirrored controller code (retires register TS-22 + UN-5).
MotionInterpreter (decomp-verbatim, Ghidra/ACE-cross-checked):
- adjust_motion (0x00528010): WalkBackward->WalkForward (x-0.649999976),
TurnLeft->TurnRight (x-1), SideStepLeft->SideStepRight (x-1), sidestep scale
0.5*(WalkAnimSpeed/SidestepAnimSpeed), RunForward early-return, per-channel
holdkey fallback to CurrentHoldKey.
- apply_run_to_command (0x00527be0): WalkForward->RunForward when speed>0 +
UNCONDITIONAL *runRate; TurnRight *1.5; SideStepRight *runRate clamp +/-3.0.
- apply_raw_movement (0x005287e0): copy raw->interpreted, adjust the 3 channels
with per-channel hold keys.
- WalkAnimSpeed unified to the retail-exact 3.11999989f (was 3.12f).
PlayerMovementController: build ONE RawMotionState from MovementInput at
forward_speed=1.0 (apply_run_to_command applies the run rate — a pre-scaled
speed would double-scale), run apply_raw_movement, then take grounded + jump
velocity straight from get_state_velocity (now correct for all directions) and
drive the keyboard turn omega from the interpreted turn_speed (BaseTurnRate x
turn_speed = the same pi/2 formula the remote path uses; numerically identical
to the old TurnRateFor). Deleted the hand-mirrored backward(x-0.65) /
strafe(x+-1.25) formulas in both the grounded and jump blocks. Mouse turn and
the auto-walk path are untouched.
Retail-faithful behavior changes (confirm in smoke): strafe is now
1.25 * ~1.248 * runRate = ~1.56*runRate, clamped via SideStepSpeed<=3.0 so
|v.X|<=3.75 (was 1.25*runRate, no scale/clamp); backward is unchanged
(3.12*0.65*runRate); backward/strafe-left velocity now comes from the pipeline
instead of returning zero. Forward run pace unchanged (4.0*runRate).
Tests: MotionNormalizationTests (17, adjust_motion/apply_run_to_command) +
MotionVelocityPipelineTests (10, apply_raw_movement->get_state_velocity golden
velocities incl. the strafe clamp + turn speeds). Full suite green (3255).
Register: TS-22 + UN-5 deleted (retired); TS-34 added (adjust_motion creature
guard is a no-op — IWeenieObject has no IsCreature; correct for the only caller,
the player). Pseudocode: docs/research/2026-07-01-d6-motion-interp-pseudocode.md.
NEXT D6.2b: the echo-gated wire forward_speed=1.0 question (evidence suggests
ACE relays rather than recomputes; verified via the smoke echo).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Ports the three retail outbound-movement packers verbatim (decomp-derived
golden bytes, confirmed via the Ghidra bridge, cross-checked vs holtburger):
- D1 — RawMotionState::Pack (0x0051ed10): new AcDream.Core.Physics.RawMotionState
data type (11 fields + actions, retail defaults) + RawMotionStatePacker that
sets a flag bit only when the field DIFFERS from its default. MoveToState.Build
now takes a RawMotionState instead of presence-based nullable params, so the
over-sent forwardSpeed=1.0 / currentHoldKey=None / default per-axis holdkeys are
no longer emitted. num_actions packs into bits 11-15 (not "bits 11-31").
- D3 — MoveToStatePack::Pack (0x005168f0) trailing byte =
(standingLongjump ? 0x02 : 0) | (contact ? 0x01 : 0); explicit contact/
standingLongjump params (standingLongjump=false honestly until the feature lands).
- D4 — JumpAction rewritten to retail JumpPack::Pack (0x00516d10): extent,
velocity, full Position, four u16 timestamps, align. Removed the spurious
objectGuid/spellId u32s; Position is now packed (it was absent). Body 56 bytes.
- Position::Pack (0x005a9640) / Frame::Pack (0x00535130) verified already-correct
(cellId, origin xyz, quaternion wxyz); locked with a golden test, no change.
GameWindow callers adapted minimally: build the RawMotionState from the existing
MovementResult values (behavior preserved except the intended D1 omissions) and
pass cellId/position/rotation to the Jump send. Pre-existing MotionInterpreter
placeholder struct RawMotionState renamed LegacyRawMotionState (D6/Phase-2 scope,
pure rename) to free the name for the retail-faithful type.
D5 audit: confirmed a real divergence — retail SendMovementEvent (0x006b4680)
stamps only last_sent_position_time after an MTS while SendPositionEvent
(0x006b4770) stamps all three; acdream's NotePositionSent stamps all three on
both paths. Left unchanged (comments added at both call sites), recorded as
register TS-33, deferred to a dedicated cadence-port slice.
Tests: RawMotionStatePackTests / MoveToStateGoldenTests / JumpActionTests /
PositionPackTests + updated MoveToStateTests / AutonomousPositionTests. Full
suite green (Core.Net.Tests 372, full solution 3228 passed / 4 pre-existing skips).
Register: TS-24/TS-25 refreshed (packer now supports actions/style; runtime
emission still deferred), TS-33 added. Roadmap L.2b shipped note added.
Spec: docs/superpowers/specs/2026-06-30-movement-wire-parity-design.md (2-6)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The register's UN-2 row recorded a contradiction: the GetMaxSpeed XML doc
claimed the bare run rate was retail-correct (~5.9 m/s catch-up, calling
the xRunAnimSpeed multiply a misread), while the implementation multiplied
by RunAnimSpeed citing ACE. Settled against the binary, not the pseudo-C:
- BN pseudo-C (acclient_2013_pseudo_c.txt:305127) renders get_max_speed as
void with a bare `this->my_run_rate;` because it DROPS x87 instructions.
- Disassembling the PDB-matched v11.4186 binary at VA 0x00527cb0: all THREE
return paths end `fld <rate>; fmul dword ptr [0x007C8918]; ret`, and the
.rdata dword at 0x007C8918 is 4.0f. Sibling get_adjusted_max_speed
(0x00527d00) carries the same trailing fmul. Verifier committed at
tools/verify_un2_fmul.py (PE parse + byte decode, rerunnable).
- Retail paths: weenie null -> 1.0 x4; InqRunRate ok -> queried x4;
InqRunRate failed -> my_run_rate x4. ACE MotionInterp.cs:665-676 matches.
Changes:
- Doc-comment rewritten: the implementation is retail-correct; the catch-up
speed 2 x get_max_speed ~= 23.5 m/s at run 200 IS retail. The 1-Hz
remote-blip symptom the old comment attributed to this multiply is
therefore UNEXPLAINED by it (if it recurs: #41 family, not this).
- Weenie-null path aligned to retail's LITERAL 1.0 default (was MyRunRate).
- Tests re-pinned to the three retail paths (the old NoWeenie test pinned
the non-retail fallback).
- Register: UN-2 row deleted per the retire rule (6 -> 5 UN rows);
shortlist renumbered.
This is the 2nd confirmed instance of the BN x87-dropout artifact class
(memory: feedback_bn_decomp_field_names) deciding a register row.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Replaces the hybrid that double-counted forward translation:
predicted body.Velocity (set per-tick by apply_current_movement) +
the seqVel-derived offset both pushed the remote body forward at
~11.7 m/s × dt for run, summing to ~23.4 m/s × dt — the user's
"way too fast" + 1-Hz blip.
Per the named-retail decomp investigation 2026-05-03 (research agent
report dispatched against acclient_2013_pseudo_c.txt for
CSequence::update + UpdatePositionInternal + UpdateObjectInternal +
adjust_offset, line citations in the env-var path comments):
CPhysicsObj::UpdateObjectInternal (0x005156b0)
→ UpdatePositionInternal (0x00512c30)
→ CPartArray::Update (writes anim root motion into the offset frame)
→ PositionManager::adjust_offset (REPLACES the offset with catch-up
when the body is far from the queue head; otherwise leaves the
anim root motion alone — Frame::operator=(arg2, &__return)
semantics, NOT additive)
→ Frame::combine (out = m_position + offset)
→ UpdatePhysicsInternal (out += body.Velocity × dt + 0.5·accel·dt²)
For a remote in steady-state RunForward where the server hasn't pushed
an explicit velocity, m_velocityVector ≈ 0 and ALL per-tick translation
comes from the animation root motion (CSequence::update_internal +
Frame::combine of crossed pos_frames keyframes). Our port doesn't
extract per-keyframe pos_frames from the .anm assets; instead
AnimationSequencer.CurrentVelocity is the synthesized equivalent
(RunAnimSpeed × ForwardSpeed averaged), passed through
PositionManager.ComputeOffset.
Concrete changes in the env-var (ACDREAM_INTERP_MANAGER=1) path:
* Pass seqVel = ae.Sequencer.CurrentVelocity to ComputeOffset (was
Vector3.Zero — that disabled the animation-root-motion source and
left only the queue catch-up to drive translation, which lagged
server pace).
* Clear rm.Body.Velocity to Vector3.Zero for grounded remotes each
tick. Mirrors retail's m_velocityVector ≈ 0 for remotes; prevents
UpdatePhysicsInternal from adding a second 11.7 m/s × dt on top of
the seqVel-driven translation.
* Stop calling apply_current_movement per tick. Retail only calls it
on motion-state changes (per cdb traces from the L.5 investigation),
not per physics tick. body.Velocity-based translation is now the
AIRBORNE-only path (gravity integration during jumps).
Also reverts an unacceptable "scaling hack" (per-tick body.Velocity
scaled by observed serverSpeed/predictedSpeed) the user explicitly
rejected as patching over an unsolved structural problem.
GetMaxSpeed reverted to RunAnimSpeed × rate (matches ACE
MotionInterp.cs:670-678; the earlier "return bare rate" change came
from a misread of an x87-decompiled get_max_speed where Binary Ninja
showed the return type as void).
AnimationSequencer.SetCycle now ALWAYS overwrites CurrentVelocity for
known locomotion cycles (Walk/WalkBackward/Run/SideStepRight/
SideStepLeft) instead of gating on `CurrentVelocity.LengthSquared() <
1e-9f`. The gate was correct for non-locomotion entities with
dat-baked HasVelocity, but for Humanoid where the dat is silent and
the only thing that could set CurrentVelocity before synthesis was a
transition link's HasVelocity flag, the gate would silently leave the
body advancing at the link's velocity instead of the cycle's intended
steady-state.
Adds wire-arrival diagnostics gated on ACDREAM_REMOTE_VEL_DIAG=1
(SETCYCLE, FWD_WIRE) used to trace the bug to ground truth.
User-confirmed improvements vs prior state:
- Steady-state run no longer "way too fast"
- Run-in-circles smoother (rectangle effect gone)
- Jump landing in correct location
- Turn-left visibly turns left
Outstanding (not addressed by this commit, deferred for next
investigation): walk↔run direct transitions don't visibly switch the
animation cycle until the next motion event fires. Both legacy and
new paths exhibit the same behavior, so the bug lives in the
SetCycle queue manipulation pipeline shared by both — not in the
per-tick translation path that this commit revises. Wire trace
confirms ACE delivers the WalkForward → RunForward transition
correctly and SetCycle does fire for it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Multi-bug fix for the env-var-gated retail-faithful remote tick path
(ACDREAM_INTERP_MANAGER=1). Combines four previously-stacked defects
into one coherent rewrite:
1. PositionManager.ComputeOffset was additive (rootMotion + correction).
Retail's PositionManager::adjust_offset (acclient @ 0x00555190 →
InterpolationManager::adjust_offset @ 0x00555d30) REPLACES the
offset frame via Frame::operator=(arg2, &__return) when catch-up
engages — it does NOT add to the rootOffset that CPartArray::Update
wrote. Switched to "correction overrides root motion" semantics.
2. MotionInterpreter.GetMaxSpeed was returning RunAnimSpeed × rate
(~11.7 m/s for run skill 200). The retail decomp at
acclient_2013_pseudo_c.txt:305127 shows get_max_speed returns the
bare run rate (~2.94) — the function's float return rides the x87
FPU stack, which Binary Ninja shows as void. Caller multiplies by
2.0 to get the catch-up speed. With the wrong return our catch-up
was 23.5 m/s instead of retail's 5.88 m/s — the queue would walk
the body 4× too aggressively.
3. The env-var TickAnimations branch was DOUBLE-COUNTING forward
translation: it applied seqVel × dt via PositionManager.ComputeOffset
AND let UpdatePhysicsInternal advance body.Position += body.Velocity
× dt. Both were ~11.7 m/s for run, so body raced at 23.4 m/s —
"way too fast" per the user. Pass seqVel=Vector3.Zero to
ComputeOffset; let body.Velocity (refreshed per tick by
apply_current_movement) drive the bulk translation alone.
4. Body orientation only applied sequencer.CurrentOmega per tick. For
the running-in-circles case ACE broadcasts ForwardCommand=RunForward
AND TurnCommand=TurnLeft on the same UpdateMotion; the sequencer
picks the RunForward cycle whose synthesized CurrentOmega is zero,
so body never rotated between UPs and body.Velocity stayed in an
out-of-date world direction — the visible "rectangle when running
circles" effect. Prefer ObservedOmega (set explicitly in
OnLiveMotionUpdated from the wire's TurnCommand + signed TurnSpeed)
when present; fall back to seqOmega for standalone turn cycles.
Also adds:
- Sequencer-reset call in the env-var landing-fallback so the legs
un-fold from Falling on land (mirrors the legacy K-fix17 path).
- LastServerZ now only updates on IsGrounded UPs, so the per-tick
landing-fallback floor doesn't drift up to the player's airborne
peak Z and force-land mid-arc — fixes the user-reported "small
landing in the air before landing on the ground" when jumping
while moving.
- VEL_DIAG now samples at UP arrival with overlapping windows, plus
TURN_WIRE / OMEGA_DIAG / FWD_WIRE diagnostics gated on
ACDREAM_REMOTE_VEL_DIAG=1 used to trace these bugs to ground truth.
Verified via live retail-driven character observation 2026-05-03:
turn-left now rotates left (was animating right with snap), running
in circles is much smoother, jumping lands on ground (no mid-air
pause). Residual ~20% steady-state overshoot for walk remains —
WalkAnimSpeed=3.12 (decompiled retail constant) doesn't match ACE's
actual broadcast walk pace (~2.6 m/s). Tracked separately.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Ports retail's CMotionInterp::get_max_speed (0x00527cb0). Returns
motion-table-derived max speed (m/s) for InterpretedState.ForwardCommand:
- RunForward: RunAnimSpeed (4.0) × (InqRunRate ?? MyRunRate)
- WalkForward: WalkAnimSpeed (3.12)
- WalkBackward: WalkAnimSpeed × 0.65 (BackwardsFactor from adjust_motion @ 0x00528010)
- otherwise: 0
Decomp note: Binary Ninja emits a spurious void return for x87 FPU-returning
functions; the actual float return is confirmed by both callers
(StickyManager::adjust_offset @ 0x00555430,
InterpolationManager::AdjustOffset @ 0x00555d52) which multiply the result
by 2.0 to produce a catch-up speed in m/s. The per-command switch is
consistent with get_state_velocity (0x00527d50) which uses the same constants.
Used by InterpolationManager.AdjustOffset in Task 5 as 2 × GetMaxSpeed().
Until Task 5 wires it, the method is unused — covered by 4 unit tests.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The decompiled get_state_velocity (FUN_00528960) literally computes
`RunAnimSpeed * ForwardSpeed` — a 4.0 × runRate world velocity. That
matches retail only when the character's MotionTable happens to bake
MotionData.Velocity.Y = 4.0 on RunForward (true for Humanoid, not
necessarily for other creatures or swapped weapon-style cycles).
When MotionData.Velocity ≠ RunAnimSpeed, the body's world velocity
drifts away from the animation's baked-in root-motion velocity, and
you see the classic "legs cycle too slowly for how fast the body is
sliding" visual bug. User reports ~30% discrepancy ("running animation
is too slow"), consistent with Humanoid RunForward's actual dat
Velocity being ~3.0 rather than the 4.0 constant.
The fix per r03 §1.3: physics body velocity = MotionData.Velocity ×
speedMod. That's exactly what AnimationSequencer.CurrentVelocity
already exposes. Route it into MotionInterpreter via an opt-in
Func<Vector3> accessor. When wired, get_state_velocity uses the
sequencer's cycle velocity as the primary forward-axis drive; when
unwired (tests, physics bodies without a sequencer), falls back to
the decompiled constant path — byte-compatible with retail on the
shapes where it actually matters.
The RunAnimSpeed × rate max-speed clamp at the bottom of
FUN_00528960 stays intact — Option B only replaces the *drive*, not
the clamp. 20 m/s phantom MotionData can't teleport the player.
Wiring: GameWindow attaches `playerAE.Sequencer.CurrentVelocity` to
`_playerController` on Tab-player-mode entry. The sequencer is always
built before the player enters chase mode, so timing is safe.
Sidestep continues to use SidestepAnimSpeed — the sequencer only
tracks the current forward cycle, so strafe is a separate axis.
6 new MotionInterpreterTests verify: accessor overrides constant path,
zero Y falls back to constant (link transitions), clamp still applies,
Ready state doesn't leak accessor value, sidestep axis is untouched.
All 717 tests green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The actual retail behavior for jump animations is a plain SubState swap —
NOT an Action overlay as I'd initially guessed. MotionCommand.Falling
(0x40000015) is a SubState cycle whose motion-table entries handle the
whole jump lifecycle:
- Links[(stance, RunForward)][Falling] = leap-into-air link
- Cycles[(stance, Falling)] = airborne cycle (loops)
- Links[(stance, Falling)][Ready/...] = landing link back to normal
Empirical verification from the diagnostic dump:
Links[0x003D0007] has 3 inner entries:
inner key: 0x41000003 (Ready)
inner key: 0x45000005 (WalkForward)
inner key: 0x40000015 (Falling) ← jackpot
SetCycle() already handles SubState + Links + Cycles resolution correctly,
so the whole fix is three lines:
if (!result.IsOnGround)
animCommand = MotionCommand.Falling;
What's in this commit:
- Added MotionCommand.Falling (0x40000015) constant + comments explaining
the retail jump-is-a-SubState flow
- GameWindow.UpdatePlayerAnimation swaps to Falling when airborne (the
cleanest possible implementation — motion table does all the work)
- Kept AnimationSequencer.PlayAction infrastructure (ported via Links
fallback + Modifiers fallback). Not needed for jump, but perfectly
valid for emotes like /wave, /bow (found in the same Links dict as
inner keys 0x13000080-0x13000083) and eventual combat attacks
- Kept MotionCommand.Jump / Jumpup / FallDown constants (unused for now
but useful reference if non-humanoid motion tables use them)
- Removed all diagnostic logging
What was learned (for future motion work):
- Retail's MotionTable.Cycles dict holds SubState loops (Ready, Walk,
Run, Falling, Crouch, etc.) by (style<<16) | (motion & 0xFFFFFF)
- MotionTable.Links dict holds TRANSITIONS between motions: the OUTER
key is the (style, fromMotion) combo; the INNER key is the TARGET
motion. The stored MotionData IS the link animation played during
the transition. This is what ACE's get_link traverses.
- MotionTable.Modifiers dict holds overlay motions (mask 0x20) — rare
for humanoids, only 8 TurnRight/SideStepRight stance variants
- Actions (mask 0x10) in retail ALSO go through Links — they're
transition animations FROM current substate, not overlays. Use
PlayAction (now correctly routed to Links dict) for them.
Jump animation now works retail-faithfully for running + jumping off.
Standing-jump behavior depends on whether the player's motion table
has a Ready→Falling link; SetCycle's fallback chain should handle it
via the style-level catch-all if the direct link is absent.
470 tests pass. Build clean.
Adds AnimationSequencer.PlayAction as the proper path for Action and
Modifier-class motions (the MotionTable.Modifiers dict, distinct from
Cycles). Action nodes are inserted before the looping cyclic tail so
they drain once and the cycle resumes naturally — leveraging the
sequencer's existing "non-looping head drains, cyclic tail wraps"
queue semantics.
What this does:
- New AnimationSequencer.PlayAction(motionCommand, speedMod=1f):
- Resolves (style<<16) | (motion&0xFFFFFF) from MotionTable.Modifiers
- Falls back to (motion&0xFFFFFF) plain key
- Silent no-op when not found (some motion tables lack these)
- Inserts AnimNodes before _firstCyclic; re-points the cursor when on
the cyclic tail so the action plays immediately
- New MotionCommand.Jump (0x2500003B) + MotionCommand.FallDown (0x10000050)
constants.
- GameWindow.UpdatePlayerAnimation fires PlayAction(Jump) on
result.JumpExtent.HasValue and PlayAction(FallDown) on JustLanded.
Key research finding: retail does NOT animate jumps.
- ACE Player.HandleActionJump explicitly clears PendingMotions and sets
IsAnimating=false during a jump (Player.cs:914-915).
- Empirical verification: the player humanoid's MotionTable only has 8
Modifier entries — all TurnRight/SideStepRight stance variants. No
Jump (0x2500003B) or FallDown (0x10000050) entries.
- Jump is a physics-only action: the character keeps whatever cycle
was active (walk/run/idle) while the physics body arcs through the
air. There is no "raise arms to jump" pose in retail.
PlayAction is still called on jump/land as a safety hatch for creature
Setups that DO carry leap animations in their Modifiers dict (drudge
jumps, monster pounces, etc.). For player humanoids it's a no-op. The
infrastructure is also ready for future emote/combat actions that
legitimately use the Modifiers dict.
470 tests pass, build clean.
Two fixes for jump physics:
- Skip ground-snap when velocity Z > 0 (prevents immediate re-landing
at high framerates where per-frame Z delta < 0.05 snap threshold)
- Guard apply_current_movement velocity write behind OnWalkable check
(prevents MotionInterpreter.DoMotion from zeroing jump velocity on
every frame while airborne)
- Guard PlayerMovementController velocity replacement behind OnWalkable
(preserves momentum during airborne flight)
Jump works locally but server packet not yet sent (BUG-002).
Facing direction mismatch logged as BUG-003.
RunRate not verified as BUG-004.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Erik