408c8e8f34
Analysis only. A candidate fix was written, built clean and passed the
three existing settler tests, then deliberately REVERTED — the only test
that discriminates it needs an EnvCell fixture that was not safe to
assemble at the end of this session. The production tree is unchanged.
Headline: the issue's framing is half wrong, and the half it misses is the
whole fix. PhysicsBody.Position's ordinary setter already carries the world
displacement into the landblock-relative frame AND calls
LandDefs.AdjustToOutside, which recomputes the outdoor cell index across
24 m cell crossings and wraps/bumps the landblock across 192 m boundaries.
So for an outdoor->outdoor settle, discarding settle.CellId costs nothing.
The live defect is EnvCells. An EnvCell id is not derivable from a world
position, and AdjustToOutside's guard ((cell & 0xFFFF) is >= 1 and <= 0x40)
deliberately excludes EnvCell ids from that path. settle.CellId is the ONLY
carrier of an EnvCell identity, and it is exactly what the settler drops —
so the defect is the issue's parenthetical ("outdoor/EnvCell seam, stacked
EnvCells"), not its main clause. That also means the fix is a no-op on the
outdoor path that dominates production and corrective only at the indoor
seam.
Recorded so the next reader does not repeat the misreading I made:
CommitTransitionPosition looks like it pairs a new cell with a stale local
origin, but line 259's `Position = worldPosition` runs the ordinary setter
first, so line 265 reads the already-updated origin. Retail anchor
CPhysicsObj::SetPositionInternal(CTransition const*) 0x00515330 commits both
objcell_id and frame, including EnvCells.
Also recorded: the three existing settler tests build bodies with no
CellPosition, so every one passes identically with or without the fix.
Shipping against them would repeat C5b finding D3 — a test that passed with
its own change reverted. The doc carries the exact discriminating test, its
required sabotage, and the fixture risk (the resolver must genuinely report
the EnvCell in settle.CellId; a fixture that silently resolves outdoor would
be green and prove nothing).
Open and unverified: whether the remote spawn-seed caller's body carries a
CellPosition at all. CommitTransitionPosition early-returns on a zero cell,
so the fix would be an inert no-op there and #276 would stay open for
remotes. Must be settled before claiming the fix closes both halves.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
114 lines
6 KiB
Markdown
114 lines
6 KiB
Markdown
# #276 remainder — SpawnPlacementSettler settle-cell discard: scoping (2026-08-06)
|
|
|
|
Analysis complete; **no production change committed.** A candidate fix was
|
|
written, built clean, and passed the three existing settler tests — then
|
|
**deliberately reverted**, because the only test that would *discriminate* it
|
|
needs an EnvCell fixture that was not safe to assemble in the session that
|
|
found this. See §4.
|
|
|
|
Written at HEAD `1d2d4bb8`.
|
|
|
|
---
|
|
|
|
## 1. The headline finding — the issue's framing is half right, and the half it misses is the whole fix
|
|
|
|
`docs/ISSUES.md:1640` says the settler "commits `settle.Position` but never
|
|
reads `settle.CellId`", so a settle crossing a cell boundary "leaves the body's
|
|
cell at the placement cell". The symptom is real. The **outdoor** half of the
|
|
explanation is not.
|
|
|
|
`PhysicsBody.Position`'s ordinary setter already carries the world displacement
|
|
into the landblock-relative frame **and** calls `LandDefs.AdjustToOutside`,
|
|
which recomputes the outdoor cell index from the local origin across 24 m cell
|
|
crossings and wraps/bumps the landblock across 192 m boundaries
|
|
(`src/AcDream.Core/Physics/PhysicsBody.cs:279-282`). So for an outdoor→outdoor
|
|
settle, `body.Position = settle.Position` **already lands the correct cell**.
|
|
Discarding `settle.CellId` costs nothing there.
|
|
|
|
**The real gap is EnvCells.** An EnvCell id is not derivable from a world
|
|
position — there is nothing for `AdjustToOutside` to recompute, and its guard
|
|
(`(cell & 0xFFFF) is >= 1 and <= 0x40`, `PhysicsBody.cs:266`) deliberately
|
|
excludes EnvCell ids (≥ 0x100) from that path entirely. The resolver's
|
|
`settle.CellId` is the *only* carrier of an EnvCell identity, and it is exactly
|
|
what the settler drops. So the live defect is the issue's parenthetical —
|
|
"outdoor/EnvCell seam, stacked EnvCells" — and not its main clause.
|
|
|
|
This matters for the fix's risk profile: the change is a **no-op on the outdoor
|
|
path that dominates production**, and corrective only at the indoor seam.
|
|
|
|
## 2. The fix (validated, then reverted)
|
|
|
|
Replace `body.Position = settle.Position;`
|
|
(`src/AcDream.Core/Physics/SpawnPlacementSettler.cs:61`) with the identical
|
|
guarded shape the per-tick resolve writeback uses at
|
|
`src/AcDream.Runtime/Physics/RuntimeOrdinaryPhysicsUpdater.cs:162-167`:
|
|
|
|
```csharp
|
|
uint resolvedCellId = settle.CellId != 0 ? settle.CellId : cellId;
|
|
body.CommitTransitionPosition(resolvedCellId, settle.Position);
|
|
```
|
|
|
|
This is the issue's own prescribed shape ("commit the settle's resolved cell
|
|
through the same body/cell channel the per-tick resolve writeback uses").
|
|
|
|
**Why the ordering is safe** — worth recording, because reading
|
|
`CommitTransitionPosition` (`PhysicsBody.cs:257-277`) in isolation suggests it
|
|
pairs the *new* cell with a *stale* local origin. It does not: line 259's
|
|
`Position = worldPosition` runs the ordinary setter first, which updates
|
|
`CellPosition.Frame.Origin`; line 265 then reads the **already-updated** origin.
|
|
Retail anchor: `CPhysicsObj::SetPositionInternal(CTransition const*)`
|
|
@0x00515330 commits both `sphere_path.curr_pos.objcell_id` and its frame,
|
|
including EnvCells.
|
|
|
|
**Verified:** `dotnet build src/AcDream.Core -c Release` → 0 errors, 0 warnings;
|
|
`SpawnPlacementSettlerTests` 3/3 pass.
|
|
|
|
## 3. Reach — one caller confirmed, one unverified
|
|
|
|
| caller | passes as `cellId` | body carries a cell? |
|
|
|---|---|---|
|
|
| `RuntimeLocalPlayerPhysicsPublicationState.cs:812` (C3c local first-entry) | `activation.Body.CellPosition.ObjCellId` | **Yes**, by construction |
|
|
| `LiveEntityNetworkUpdateController.cs:297` (remote spawn seed, #270) | a `cellId` parameter | **UNVERIFIED** |
|
|
|
|
`CommitTransitionPosition` early-returns when `CellPosition.ObjCellId == 0`
|
|
(`PhysicsBody.cs:261`), so on any body without a cell the fix is an inert no-op
|
|
— safe, but it would mean #276 stays open for remotes even after the change.
|
|
**Establish the remote body's `CellPosition` provenance before claiming the fix
|
|
closes both halves.** The register row AD-61 covers this shared settler.
|
|
|
|
## 4. Why no test was committed, and the exact test to write
|
|
|
|
The three existing tests
|
|
(`tests/AcDream.Core.Tests/Physics/SpawnPlacementSettlerTests.cs`) construct
|
|
bodies with **no `CellPosition`**, so every one of them passes identically with
|
|
or without the fix — they cannot discriminate it. Shipping the change against
|
|
them would repeat the defect this campaign already paid for once (a
|
|
conservation test that passed with its own change reverted, C5b finding D3).
|
|
|
|
Per §1 the discriminating case must be **indoor**, because an outdoor fixture
|
|
also passes both ways via `AdjustToOutside`. Required shape:
|
|
|
|
1. `AddLandblock(0xA9B40000, flat terrain, [envCellSurface], portalPlanes, …)`
|
|
with an EnvCell surface id ≥ `0x100` — build the floor with
|
|
`CellSurfaceTests.MakeFlatFloor`'s 4-vertex quad pattern
|
|
(`tests/AcDream.Core.Tests/Physics/CellSurfaceTests.cs:20-38`).
|
|
2. Body seeded with `CellPosition.ObjCellId` = an **outdoor** placement cell
|
|
(e.g. `0xA9B40001`), positioned just above the EnvCell floor.
|
|
3. Call `TrySettle` with that outdoor cell as the wire `cellId`.
|
|
4. **Assert** `body.CellPosition.ObjCellId == 0xA9B40100` (the EnvCell).
|
|
5. **Sabotage that must redden it:** restore `body.Position = settle.Position`
|
|
— the cell stays outdoor, because `AdjustToOutside` cannot invent an EnvCell.
|
|
|
|
The open risk, and the reason this was not attempted at the end of a long
|
|
session: step 1 must produce a fixture where the resolver genuinely reports the
|
|
EnvCell in `settle.CellId`. That may require portal planes for cell membership
|
|
(`CellTransit.FindCellList`). A fixture that silently resolves to the outdoor
|
|
cell would make the test pass for the wrong reason in the *other* direction —
|
|
green, and proving nothing. Verify `settle.CellId` is the EnvCell id before
|
|
trusting the assertion.
|
|
|
|
## 5. Recommended next step
|
|
|
|
Write the §4 test first, watch it fail against HEAD, then apply §2 and watch it
|
|
pass, then sabotage. One commit, ~40 lines of test, ~8 of production. Resolve §3
|
|
in the same pass or state plainly that the remote half remains open.
|