erik/acdream
1
0
Fork 0
Code Issues Pull requests Projects Releases 6 Packages Wiki Activity Actions

fix(physics): preserve retail step-down probe state

Browse source
This commit is contained in:
Erik 2026-07-31 13:24:25 +02:00
parent 1fd5da67b4
commit acec33eca8
5 changed files with 249 additions and 36 deletions
Download patch file Download diff file Expand all files Collapse all files

2
docs/architecture/retail-divergence-register.md
View file

@ -130,7 +130,7 @@ AP-94..AP-112 for the confirmed retail-UI completion gaps.
| ~~AP-3~~ | **RETIRED 2026-07-31 (Campaign P Slice 1B).** `TransitionalInsert` now returns `OK_TS` immediately for every valid contact plane. Its ordinary StepDown tail is reachable only from invalid contact and retains the retail Contact / `!sphere_path.step_down` / check-cell / ObjectInfo.StepDown gates plus the exact one-versus-two-sphere probe split. | `src/AcDream.Core/Physics/TransitionTypes.cs` (`TransitionalInsert`, `GetStepDownProbePlan`); `tests/AcDream.Core.Tests/Physics/RetailEdgeResponseOrderingTests.cs` | — | — | `CTransition::transitional_insert` 0x0050B6F0, named-retail pseudo-C pc:273191–273307 |
| ~~AP-4~~ | **RETIRED 2026-07-31 (Campaign P Slice 1B).** `EdgeSlideAfterStepDownFailed` now evaluates retail Branch 1 (`!OnWalkable || !EdgeSlide` → restore + `OK_TS`) before the steep-contact `CliffSlide` branch. The former compensation is removed. | `src/AcDream.Core/Physics/TransitionTypes.cs` (`EdgeSlideAfterStepDownFailed`); `tests/AcDream.Core.Tests/Physics/RetailEdgeResponseOrderingTests.cs` | — | — | `CTransition::edge_slide` 0x0050B3D0, named-retail pseudo-C pc:273001–273090 |
| ~~AP-5~~ | **RETIRED 2026-07-31 (Campaign P Slice 2A).** `DoStepDown` no longer accepts a caller-controlled `runPlacement` bypass. After the transitional support probe and retail `check_walkables` gate succeed, ordinary contact maintenance, edge-slide back-probes, and StepUp all switch to `PLACEMENT_INSERT`, reset `walk_interp` to 1, run the final insertion, restore the prior insert type, and accept only `OK_TS`. The former wall-slide justification is addressed at the actual placement dispatcher boundary: its retail epsilon-shaved overlap test permits exact wall tangency but rejects real penetration; no StepDown path skips validation. | `src/AcDream.Core/Physics/TransitionTypes.cs` (`DoStepDown`); `src/AcDream.Core/Physics/BSPQuery.cs` / `FlatBspQuery.cs` (Placement dispatcher); `tests/AcDream.Core.Tests/Physics/RetailStepDownPlacementTests.cs` | — | — | `CTransition::step_down` 0x0050B2A0 pc:272946–272998; `BSPTREE::find_collisions` Placement branch 0x0053A440 pc:323742; `CSphere::intersects_sphere` 0x00537A80; `CCylSphere::intersects_sphere` 0x0053B440 |
| ~~AP-5~~ | **RETIRED 2026-07-31 (Campaign P Slice 2A).** `DoStepDown` no longer accepts a caller-controlled `runPlacement` bypass. It resets `walk_interp` once at entry; after the transitional support probe and retail `check_walkables` gate succeed, ordinary contact maintenance, edge-slide back-probes, and StepUp all switch to `PLACEMENT_INSERT` with the exact carried interpolation value, run the final insertion, restore the prior insert type, and accept only `OK_TS`. Nested `DoCheckWalkable` uses local current-position saves and preserves the outer `SPHEREPATH` backup pair needed by edge-slide. The former wall-slide justification is addressed at the actual placement dispatcher boundary: its retail epsilon-shaved overlap test permits exact wall tangency but rejects real penetration; no StepDown path skips validation. | `src/AcDream.Core/Physics/TransitionTypes.cs` (`DoStepDown`, `DoCheckWalkable`); `src/AcDream.Core/Physics/BSPQuery.cs` / `FlatBspQuery.cs` (Placement dispatcher); `tests/AcDream.Core.Tests/Physics/RetailStepDownPlacementTests.cs`; `tests/AcDream.Core.Tests/Physics/RetailEdgeResponseOrderingTests.cs` | — | — | `CTransition::check_walkable` 0x0050AFF0 pc:272811–272856; `CTransition::step_down` 0x0050B2A0 pc:272946–272998; `BSPTREE::find_collisions` Placement branch 0x0053A440 pc:323742; `CSphere::intersects_sphere` 0x00537A80; `CCylSphere::intersects_sphere` 0x0053B440 |
| ~~AP-7~~ | **RETIRED 2026-07-30 (Campaign P Slice P2) — the "state gate" was a BN decompiler artifact, not a locomotion exemption.** `calc_friction` now ports retail's confirmed 0.25f threshold (`if (angle >= 0.25f) return;`) unconditionally, no special-cased gate. The "state check at pc:276702" the old row cited is `PhysicsState.Sledding` (confirmed via ACE's `PhysicsObj.calc_friction`, references/ACE/Source/ACE.Server/Physics/PhysicsObj.cs:2120-2141, and `SLEDDING_PS=0x800000` in acclient.h:2838) — it gates the 1.5625/6.25/near-flat friction-value OVERRIDE, not the threshold return itself; acdream had no live Sledding setter then or now (see #166 research, docs/research/2026-07-30-response-layer-edge-family-pseudocode.md §3), so the branch was simply unreachable dead code, not an exemption for ordinary walking. The reverted 2026-04-30 L.3c attempt (naive 0.0→0.25 bump, forward locomotion 3→0.16 m/s in `PlayerMovementControllerTests`) does not reproduce on the production graphical local-player path post-R6: `PlayerMovementController` zeroes `Velocity.X/Y` to exactly zero every tick before `calc_friction` runs whenever animation root motion drives the walk, so friction has no horizontal velocity left to hammer (pinned at the PhysicsBody level by `GroundedRootMotion_FrictionThreshold_DoesNotHammerLocomotionTests`). The headless/`get_state_velocity` movement-controller path and remote/NPC movers still feed real velocity into this function and remain the ones to watch if a similar regression resurfaces there. **CORRECTION (2026-07-30, same day, #265/#166 capture bisect):** the sentence above undersold the gap — `calc_friction` wasn't merely "no horizontal velocity to hammer," it was structurally UNREACHABLE with meaningful data on ANY grounded path: (a) the animation-root-motion path zeroed `Velocity.X/Y` outright every tick (the actual #265/#166 root cause, ten days pre-existing, not a Campaign-P regression), and (b) `PhysicsBody.GroundNormal` — the vector `calc_friction` dots velocity against — had ZERO production writers anywhere and silently defaulted to `Vector3.UnitZ` forever, so even surviving velocity would have been tested against a fake flat-ground normal on any real slope. Both gaps are now closed: `PlayerMovementController.cs`'s grounded block no longer reconstructs `Velocity` for the animation-root-motion case, and `PhysicsEngine.cs` syncs `body.GroundNormal` from the committed `ContactPlane.Normal` at the same commit point that already publishes `ContactPlane`. The 0.25f threshold port itself (this row's original subject) was always correct — it just had nothing real to operate on until this fix. See `docs/research/2026-07-30-265-capture-bisect.md`'s as-fixed addendum. | `src/AcDream.Core/Physics/PhysicsBody.cs` (`calc_friction`); `src/AcDream.Core/Physics/PhysicsEngine.cs` (`GroundNormal` wiring); `src/AcDream.Runtime/Gameplay/PlayerMovementController.cs` (grounded-velocity fix); `tests/AcDream.Core.Tests/Physics/PhysicsBodyTests.cs` (AP-7 test block); `tests/AcDream.Core.Tests/Physics/Issue265SteepSlopeCaptureBisectTests.cs`; `tests/AcDream.Runtime.Tests/Gameplay/PlayerMovementControllerTests.cs` | — | — | `CPhysicsObj::calc_friction` pc:276694-276822 (0050ee70); ACE `PhysicsObj.calc_friction` PhysicsObj.cs:2120-2141; `docs/research/2026-07-30-response-layer-edge-family-pseudocode.md` §1; `docs/research/2026-07-30-265-capture-bisect.md` |
| ~~AP-10~~ | **RETIRED 2026-07-30 (Campaign P Slice P4) — the retail 0.1 m dry-corner water sink-in is restored.** `TerrainSurface.SampleWaterDepth` (`src/AcDream.Core/Physics/TerrainSurface.cs`) now returns 0.1 for a partially-water cell's dry corner instead of the collapsed 0. The row's own "destabilizes the touch check" justification turned out to be structurally true of retail too (a skipped `SetContactPlane` reassertion is not a fall in ANY of retail/ACE/acdream, because `Contact`/`OnWalkable` are STICKY — `PhysicsEngine.ResolveWithTransition`'s `onGround` computation ORs the fresh per-call `ContactPlaneValid` with the seeded, persistent `PhysicsBody.TransientState.OnWalkable` bit) — traced and confirmed in this slice; see `docs/research/2026-07-29-remote-and-world-specials-pseudocode.md` §5.2. `PhysicsEngine.SampleTerrainWalkable`'s `isWater = waterDepth >= 0.45f` threshold means the restore does not flip the dry corner's water classification (0.1 still < 0.45) — only the sink-in depth changes. Full Release suite green (no regression) proves the sticky-bit argument held in practice, not just in theory. | `src/AcDream.Core/Physics/TerrainSurface.cs` (`SampleWaterDepth`) | — | — | `ObjCell.get_water_depth` / `calc_water_depth` (via ACE port); `docs/research/2026-07-29-remote-and-world-specials-pseudocode.md` §5.1-5.2 |
| AP-11 | Hand-authored 4-keyframe fallback sky set (sunrise/noon/sunset, fog ~80–350 m) when the Region dat isn't loaded yet | `src/AcDream.Core/World/SkyState.cs:167` | A renderable sky is needed during boot before the Region dat parses; safety net on region-load failure | Any window where the fallback is active shows sky/fog lighting only roughly resembling retail's dat-driven values | SkyTimeOfDay keyframes, Region dat 0x13000000 |

26
docs/research/2026-07-30-response-layer-edge-family-pseudocode.md
View file

@ -1146,6 +1146,14 @@ the `EdgeSlide && !StepUp` walkable-support gate, and then always:
4. restores the saved insert type; and
5. returns true only for `OK_TS`.
The `walk_interp = 1` assignment occurs once at `step_down` entry. Retail does
not reset it again before the final Placement insertion: that insertion sees
and carries the exact interpolation value left by the support probe. Likewise,
`CTransition::check_walkable` (`0x0050AFF0`, pseudo-C:272811–272856) saves its
temporary `check_pos` and cell in stack locals. It does not overwrite
`SPHEREPATH::backup_check_pos`/`backup_cell`, because the enclosing edge-slide
still needs that outer failed candidate after the nested support probe.
This is the same tail for normal grounded contact maintenance, the
`edge_slide` current-position back-probe, and StepUp. The former acdream
`runPlacement: false` argument on the first two callers was therefore a real
@ -1161,12 +1169,16 @@ epsilon is rejected. With the response-order fixes from §8 in place, restoring
the mandatory placement tail does not reproduce the old wall stall.
`RetailStepDownPlacementTests` pins the mechanism end to end for one- and
two-sphere movers, ordinary contact maintenance, StepUp, and a supported
candidate that overlaps only during final Placement. It also compares graph
and prepared-flat Placement at exact wall tangency and beyond-epsilon overlap.
The existing multi-frame roof/edge fixtures continue to compare the complete
`ResolveResult` and persistent `PhysicsBody` state by raw bits, and the #273,
#271, #185, StepUp, transition-retry, and TS-4 controls remain unchanged. AP-5
is retired; TS-4 is intentionally untouched.
two-sphere movers, including raw-bit `walk_interp` carry, nested
`check_walkable` backup preservation, ordinary contact maintenance, StepUp,
and a supported candidate that overlaps only during final Placement. It also
compares graph and prepared-flat Placement at exact wall tangency and
beyond-epsilon overlap. The full-engine floor-plus-wall replay drives ten
grounded maintenance frames for both sphere counts, requires one Placement for
every successful support maintenance, proves no freeze or penetration plus
tangential progress, and compares complete `ResolveResult` and persistent
`PhysicsBody` state by raw bits between parsed-graph and prepared-flat paths.
The #273, #271, #185, StepUp, transition-retry, and TS-4 controls remain
unchanged. AP-5 is retired; TS-4 is intentionally untouched.
---