diag(render): flap re-diagnosed as portal-flood re-clip DRIFT; physics + camera REFUTED
The 2026-06-08 AM "physics rest micro-jitter" diagnosis is refuted with primary
evidence (door-recheck 216K standstill records: 0 position re-snaps; player
byte-stable during the flap). Two adversarial verification sub-agents confirmed:
- Retail roots the render at the camera viewer_cell (swept from the player via
SmartBox::update_viewer 0x453ce0; DrawInside(viewer_cell) 0x453aa0) and toggles
DrawInside / LScape::draw -- so acdream's eye-cell rooting + inside/outside
toggle are RETAIL-FAITHFUL. The locked-design "root at player cell" is wrong.
- The flap is render membership instability, eye-motion-driven: the visible-cell
set oscillates (8<->3) as the eye sweeps monotonically. Root = the
re-enqueue-on-growth DRIFT (PortalVisibilityBuilder.cs:322, MaxReprocessPerCell
=16) re-clipping each grown cell every round -> sub-cm eye jitter flips membership.
Fix (spec, not yet implemented): verbatim port of retail's enqueue-once flood
(ConstructView + AddViewToPortals): enqueue once on first discovery, clip each
cell's portals once, union late growth in place (AddToCell) + draw-reorder
(FixCellList), never re-enqueue. Kills the drift; rooting/camera/seal untouched.
This commit lands VERIFIED GROUNDWORK + design only:
- spec: docs/superpowers/specs/2026-06-08-portal-flood-enqueue-once-port-design.md
- findings: docs/research/2026-06-08-flap-physics-diagnosis-REFUTED-its-render-membership.md
- [pv-input] probe gains rawPlayer + yaw (disambiguates the varying input)
- 4 GREEN physics rest-stability tests (prove rest is bit-stable -> flap not physics)
- apparatus: launch-flap-capture.ps1, analyze_flap_live.py, find_burst.py
- captured fixtures: tests/.../Fixtures/flap-doorway/0xA9B4017{0..5}.json
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
parent
d6aa526dd3
commit
6c3a96b26e
14 changed files with 8231 additions and 1 deletions
|
|
@ -34,6 +34,97 @@ public class PlayerMovementControllerTests
|
|||
Assert.Equal(96f, result.Position.Y, precision: 1);
|
||||
}
|
||||
|
||||
// ── Indoor-flap root cause: resting-body bit-stability ────────────────────
|
||||
//
|
||||
// The indoor render "flap" (textures battling at the cottage doorway) is
|
||||
// portal-flood membership instability. PortalVisibilityBuilder.Build is a
|
||||
// proven-deterministic pure function, so the membership can only flip if its
|
||||
// INPUT (the camera eye, derived from the player RenderPosition) varies.
|
||||
// Live 6-dp capture (pvinput.log:54) shows the player RenderPosition carries
|
||||
// a perpetual ~1-ULP flicker at rest (Z 94.000000 <-> 93.999992 — exactly one
|
||||
// float mantissa step). ComputeRenderPosition is Vector3.Lerp(_prevPhysicsPos,
|
||||
// _currPhysicsPos, alpha), and Lerp(a, a, t) == a exactly, so a jittering
|
||||
// RenderPosition at rest means the physics body's resting Position is NOT
|
||||
// bit-stable between ticks. Retail's authoritative local position is bit-stable
|
||||
// at rest (validate_transition -> kill_velocity on every grounded contact), so
|
||||
// retail never flaps.
|
||||
//
|
||||
// This test pins the physics-side invariant: a grounded body with no input
|
||||
// must hold a byte-identical position across many frames. It PASSES — which
|
||||
// is itself the evidence: the physics resting position is bit-stable, so the
|
||||
// doorway flap is NOT a physics-rest jitter. See
|
||||
// docs/research/2026-06-08-flap-physics-diagnosis-REFUTED-its-render-membership.md
|
||||
// (the flap is render-side portal-flood membership instability at the grazing
|
||||
// doorway portal under a sweeping camera eye). Kept as a regression guard.
|
||||
[Fact]
|
||||
public void Update_AtRestNoInput_RenderPositionBitStableAcrossManyFrames()
|
||||
{
|
||||
var engine = MakeFlatEngine();
|
||||
var controller = new PlayerMovementController(engine);
|
||||
var rest = new Vector3(96f, 96f, 50f);
|
||||
controller.SetPosition(rest, 0x0001);
|
||||
|
||||
// Settle one frame so the resolver establishes its rest state, then
|
||||
// capture the baseline the body must hold.
|
||||
var settled = controller.Update(1f / 60f, new MovementInput());
|
||||
Vector3 baselineRender = settled.RenderPosition;
|
||||
Vector3 baselinePhysics = settled.Position;
|
||||
|
||||
// Hold still for ~10 s of 60 Hz frames (crosses MinQuantum every ~2
|
||||
// frames, so the 30 Hz physics tick fires throughout — same cadence as
|
||||
// live). Any deviation, even one ULP, is the flap's root cause.
|
||||
float maxRenderDev = 0f;
|
||||
float maxPhysicsDev = 0f;
|
||||
for (int i = 0; i < 600; i++)
|
||||
{
|
||||
var r = controller.Update(1f / 60f, new MovementInput());
|
||||
maxRenderDev = MathF.Max(maxRenderDev, (r.RenderPosition - baselineRender).Length());
|
||||
maxPhysicsDev = MathF.Max(maxPhysicsDev, (r.Position - baselinePhysics).Length());
|
||||
}
|
||||
|
||||
Assert.True(
|
||||
maxRenderDev == 0f && maxPhysicsDev == 0f,
|
||||
$"resting body drifted: render={maxRenderDev * 1e6f:F3} µm, " +
|
||||
$"physics={maxPhysicsDev * 1e6f:F3} µm; expected byte-identical rest");
|
||||
}
|
||||
|
||||
// After walking then releasing input, the body must SETTLE to a
|
||||
// byte-identical resting position — not keep blipping a residual velocity.
|
||||
// This models the live flap: the player walks to the cottage doorway and
|
||||
// stops, and the eye then carries a ~1-ULP jitter that flips portal-flood
|
||||
// membership. Flat-terrain variant: if even this drifts, the residual-after-
|
||||
// motion path is the root and it is not indoor-specific.
|
||||
[Fact]
|
||||
public void Update_WalkThenStop_SettlesToBitStableRest()
|
||||
{
|
||||
var engine = MakeFlatEngine();
|
||||
var controller = new PlayerMovementController(engine);
|
||||
controller.SetPosition(new Vector3(96f, 96f, 50f), 0x0001);
|
||||
controller.Yaw = 0f;
|
||||
|
||||
// Walk forward ~0.5 s, then release.
|
||||
for (int i = 0; i < 30; i++)
|
||||
controller.Update(1f / 60f, new MovementInput(Forward: true));
|
||||
// Let velocity decay / state settle.
|
||||
for (int i = 0; i < 30; i++)
|
||||
controller.Update(1f / 60f, new MovementInput());
|
||||
|
||||
var settled = controller.Update(1f / 60f, new MovementInput());
|
||||
Vector3 basePos = settled.Position;
|
||||
Vector3 baseRender = settled.RenderPosition;
|
||||
|
||||
float maxPos = 0f, maxRender = 0f;
|
||||
for (int i = 0; i < 600; i++)
|
||||
{
|
||||
var r = controller.Update(1f / 60f, new MovementInput());
|
||||
maxPos = MathF.Max(maxPos, (r.Position - basePos).Length());
|
||||
maxRender = MathF.Max(maxRender, (r.RenderPosition - baseRender).Length());
|
||||
}
|
||||
|
||||
Assert.True(maxPos == 0f && maxRender == 0f,
|
||||
$"post-walk rest drifted: pos={maxPos * 1e6f:F3} µm, render={maxRender * 1e6f:F3} µm");
|
||||
}
|
||||
|
||||
[Fact]
|
||||
public void Update_ForwardInput_MovesInFacingDirection()
|
||||
{
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue