fix(teleport): preserve dungeon cell identity on respawn (#215)

Classify teleport landblock transitions from the source and destination Position cell IDs instead of render-space coordinates. This keeps same-dungeon death respawns on their resident floor physics while preserving the existing true cross-landblock streaming path.

Co-Authored-By: Codex <noreply@openai.com>

docs/architecture/acdream-architecture.md

@@ -412,6 +412,24 @@ replace state produced by another job. CPU mesh extraction may still be schedule
 onto `ObjectMeshManager`'s thread-safe work queue, but the worker never publishes
 partially hydrated cell membership or shell placement to the renderer.
 
+### Teleport cell identity at the streaming boundary
+
+The network and physics layers carry retail's complete `Position`: an
+authoritative `objcell_id` plus a frame. Any streaming decision that asks
+whether a teleport crossed a landblock compares the source and destination cell
+IDs. It must never reconstruct either landblock by flooring render/world XYZ.
+Dungeon EnvCells have arbitrary frame origins, including valid negative local
+coordinates, so XYZ-to-grid inference can evict the landblock the player is
+still standing in.
+
+`TeleportLandblockTransition` owns this App-layer classification. When the
+source cell is unavailable before first placement, it uses the typed current
+streaming-center landblock as the fallback identity. The existing async
+recenter/hydration path runs only for a real cell-ID landblock change; a
+same-landblock teleport keeps the already-published render and physics unit.
+This is the asynchronous adapter for retail's direct full-`Position` placement
+(`SmartBox::TeleportPlayer @ 0x00453910`).
+
 ## Roadmap Model
 
 The old R1-R8 architecture sequence was a useful early refactor sketch, but it

docs/architecture/retail-divergence-register.md

@@ -146,7 +146,7 @@ AP-94..AP-112 for the confirmed retail-UI completion gaps.
 | AP-32 | Cell shells DRAW +0.02 m above the dat EnvCell origin (`ShellDrawLiftZ`, z-fight vs coplanar terrain); retail draws at the origin verbatim. Split invariant: PHYSICS + visibility graph UNLIFTED (f35cb8b, **#119**-residual), every DRAW-space consumer of portal/cell geometry LIFTED (OutsideView color gate via `Build(drawLiftZ)`, seal/punch fans — **#130**) | `src/AcDream.App/Rendering/GameWindow.cs:5604` (const at `PortalVisibilityBuilder.ShellDrawLiftZ`) | Shell floors coplanar with terrain z-fight in our z-buffered frame; the 2 cm lift is the documented stand-in | A new draw-space consumer of portal/cell polygons that forgets the lift re-opens a 2 cm seam at horizontal aperture edges (the #130 top-edge strip, ~7 px at 2.4 m); a visibility consumer that picks up the LIFTED transform re-opens the #119-residual horizontal-portal side-cull | retail draws cell geometry at the dat EnvCell origin (no lift) |
 | AP-33 | Interior-root look-in cells (**#124** sub-pass) draw their statics + DYNAMICS + emitters WHOLE — no per-part/per-object viewcone check; retail viewconeCheck's each vs the installed view (the **#131** portal closure: a server object in a look-in cell drew nowhere — dynamics-last culls cells absent from the main cone, and post-seal it z-fails anyway) | `src/AcDream.App/Rendering/RetailPViewRenderer.cs` (`DrawBuildingLookIns`) | The main viewcone has no entries for look-in cells; over-include is the safe direction (z-correct, repainted outside apertures by the root's shells); look-in cell counts are small (~1-3 cells) | A few wasted draws on content outside the doorway region (repainted); no under-draw direction remains | `viewconeCheck` 0x0054c250; nested `DrawCells` objects pc:432878 |
 | AP-34 | Landscape-stage alpha deferral is a TWO-PHASE slice split (statics-early / dynamics+particles+weather-late around the **#124** look-ins) + outdoor-root attached scene emitters moved to the post-frame pass, not retail's single deferred alpha flush. Residual: building exteriors' / outside-stage dynamics' own translucent MESH batches still draw within their stage draw call (before later stage content) | `src/AcDream.App/Rendering/RetailPViewRenderer.cs` (`DrawLandscapeThroughOutsideView` late loop) + `GameWindow` post-frame Scene pass | The MDI dispatcher draws translucency inside each Draw call; a faithful FlushAlphaList port needs a global deferred alpha list across all landscape draws — the split covers the user-visible cases (#131 portal swirl, #132 candle flame indoors + outdoors) | Translucent landscape content drawn early and screen-overlapped by content drawn later in the stage gets overpainted (no depth self-protection) — the portal-swirl/candle-flame class re-appears in the residual configurations | `D3DPolyRender::FlushAlphaList` (DrawCells pc:432722) |
-| AP-36 | Dungeon streaming gate triggers on the player's CURRENT cell being a sealed EnvCell (`CurrCell.IsEnv && !SeenOutside`), an approximation of ACE's full landblock `IsDungeon` (all-heights-zero + NumCells>0 + Buildings.Count==0). The retail BEHAVIOR (a dungeon loads no adjacent landblocks) is faithful — only the runtime TRIGGER is the cheap cell predicate instead of classifying the center landblock. **#135 pre-collapse:** at login/teleport the same collapse is triggered EARLY (the instant the streaming center is recentered onto the spawn/dest cell) via `IsSealedDungeonCell` reading the EnvCell **dat** `SeenOutside` flag — because the physics `CurrCell` is null until placement, which waits for hydration; without the early trigger the full 25×25 ocean-grid window loads then unloads (the ~30 s login FPS ramp). **#145/#138 teleport-hold suppression:** during a teleport arrival HOLD the player is unplaced, so `CurrCell` is the frozen SOURCE cell, not the destination; the gate is suppressed for the hold (`DungeonStreamingGate.Compute(isTeleportHold:true)` → not-inside-dungeon) so a teleport OUT of a dungeon follows the destination (the PortalSpace observer pin) and `ExitDungeonExpand`s, instead of re-pinning streaming onto the source dungeon (which left the outdoor destination un-hydrated → 600-frame readiness timeout → force-snap to ocean — the #145 "second teleport does nothing" + #138 incomplete-world) | `src/AcDream.App/Streaming/DungeonStreamingGate.cs` (`Compute` — per-frame predicate + teleport-hold suppression, called from `GameWindow.OnUpdate` ~:7401) + `GameWindow:IsSealedDungeonCell` + `:OnLiveEntitySpawnedLocked`/`:OnLivePositionUpdated` (login/teleport pre-collapse hooks) + `src/AcDream.App/Streaming/StreamingController.cs` (collapse/expand/`PreCollapseToDungeon`) | The predicate is already computed for sun/sky gating (playerInsideCell) and exactly matches for sealed dungeons vs windowed building interiors (SeenOutside=true → not gated); no landblock re-classification needed. The dat-flag read is the same `EnvCellFlags.SeenOutside` the hydrated `ObjCell.SeenOutside` is built from (`EnvCell.cs:72`/`PhysicsDataCache.cs:224`), so the pre-collapse decision matches the eventual per-frame gate exactly | A dungeon cell that reports SeenOutside (an entrance cell open to the surface) briefly un-collapses and re-streams the window; a hypothetical windowless building back-room (IsEnv && !SeenOutside but HasBuildings) would wrongly collapse its outdoor neighbors; a sealed-dungeon entrance cell that is itself SeenOutside is simply MISSED by the early trigger and falls back to the existing late collapse (no worse than before #135) | ACE `LandblockManager.GetAdjacentIDs` (dungeons→empty) Landblock.cs:577-582; `IsDungeon` Landblock.cs:1264-1277 |
+| AP-36 | Dungeon streaming gate triggers on the player's CURRENT cell being a sealed EnvCell (`CurrCell.IsEnv && !SeenOutside`), an approximation of ACE's full landblock `IsDungeon` (all-heights-zero + NumCells>0 + Buildings.Count==0). The retail BEHAVIOR (a dungeon loads no adjacent landblocks) is faithful — only the runtime TRIGGER is the cheap cell predicate instead of classifying the center landblock. **#135 pre-collapse:** at login/teleport the same collapse is triggered EARLY (the instant the streaming center is recentered onto the spawn/dest cell) via `IsSealedDungeonCell` reading the EnvCell **dat** `SeenOutside` flag — because the physics `CurrCell` is null until placement, which waits for hydration; without the early trigger the full 25×25 ocean-grid window loads then unloads (the ~30 s login FPS ramp). **#215 cell identity:** the pre-collapse/recenter decision compares the player's current `Position.objcell_id` landblock with the received destination `objcell_id`; it never reconstructs the source from XYZ because dungeon frame origins may be negative. **#145/#138 teleport-hold suppression:** during a teleport arrival HOLD the player is unplaced, so `CurrCell` is the frozen SOURCE cell, not the destination; the gate is suppressed for the hold (`DungeonStreamingGate.Compute(isTeleportHold:true)` → not-inside-dungeon) so a teleport OUT of a dungeon follows the destination (the PortalSpace observer pin) and `ExitDungeonExpand`s, instead of re-pinning streaming onto the source dungeon (which left the outdoor destination un-hydrated → 600-frame readiness timeout → force-snap to ocean — the #145 "second teleport does nothing" + #138 incomplete-world) | `src/AcDream.App/Streaming/TeleportLandblockTransition.cs` (source/destination cell-ID classification) + `src/AcDream.App/Streaming/DungeonStreamingGate.cs` (`Compute` — per-frame predicate + teleport-hold suppression, called from `GameWindow.OnUpdate` ~:7401) + `GameWindow:IsSealedDungeonCell` + `:OnLiveEntitySpawnedLocked`/`:OnLivePositionUpdated` (login/teleport pre-collapse hooks) + `src/AcDream.App/Streaming/StreamingController.cs` (collapse/expand/`PreCollapseToDungeon`) | The predicate is already computed for sun/sky gating (playerInsideCell) and exactly matches for sealed dungeons vs windowed building interiors (SeenOutside=true → not gated); no landblock re-classification needed. The dat-flag read is the same `EnvCellFlags.SeenOutside` the hydrated `ObjCell.SeenOutside` is built from (`EnvCell.cs:72`/`PhysicsDataCache.cs:224`), so the pre-collapse decision matches the eventual per-frame gate exactly. The cell-ID comparison matches retail's complete `Position` flow. | A dungeon cell that reports SeenOutside (an entrance cell open to the surface) briefly un-collapses and re-streams the window; a hypothetical windowless building back-room (IsEnv && !SeenOutside but HasBuildings) would wrongly collapse its outdoor neighbors; a sealed-dungeon entrance cell that is itself SeenOutside is simply MISSED by the early trigger and falls back to the existing late collapse (no worse than before #135) | ACE `LandblockManager.GetAdjacentIDs` (dungeons→empty) Landblock.cs:577-582; `IsDungeon` Landblock.cs:1264-1277; retail `SmartBox::TeleportPlayer` 0x00453910 |
 | AP-43 | Per-object torch (point/spot) lighting AND sun are both gated on the OBJECT's own cell via the same `IndoorObjectReceivesTorches(ParentCellId)` predicate (`(id & 0xFFFF) >= 0x0100`): indoor objects (EnvCell-parented) get torches + NO sun; outdoor objects get the SUN + ambient + NO torches. This is the faithful per-draw port of retail's `useSunlight` gate — `DrawMeshInternal` (0x0059f398) calls `minimize_object_lighting` only `if (Render::useSunlight == 0)`, and `PView::DrawCells` (0x005a4840) calls `useSunlightSet(1)` (0x005a485a) for the outdoor stage and `useSunlightSet(0)` (0x005a49f3) for the interior-cell stage. **#142 (2026-06-20):** the sun gate is now PER-INSTANCE in the shader (binding=6 `instanceIndoor[]` flag in `mesh_modern.vert`, filled by `AppendCurrentLightSet`) — it was previously a per-FRAME global keyed on the PLAYER cell (`UpdateSunFromSky`). The per-frame global is retained for sealed dungeons (correctly kills the sun frame-wide when no sky is visible). **Residual:** the `ebp_2` second seen-outside test in `CellManager::ChangePosition` (0x004559B0) is unaudited — unclear whether it changes the ambient/sun regime for a subset of cells. No observed behavioral impact in tested cells. | `src/AcDream.App/Rendering/Wb/WbDrawDispatcher.cs` (`IndoorObjectReceivesTorches`, `ComputeEntityLightSet`, `AppendCurrentLightSet`, `_instIndoorSsbo`/`_indoorData`/`InstanceGroup.IndoorFlags`); `src/AcDream.App/Rendering/Shaders/mesh_modern.vert` (binding=6 `instanceIndoor[]` gate on sun loop); per-frame sun `src/AcDream.App/Rendering/GameWindow.cs:10421` (`UpdateSunFromSky`, unchanged) | Torches: outdoor objects never torch-lit (exact retail). Sun: indoor objects (furniture, NPCs, player in a windowed building) never sun-lit (exact retail per-stage). Ambient: per-player-cell regime unchanged (exact retail `ChangePosition`). | The `ebp_2` unaudited test in `ChangePosition` could affect a narrow class of cells (entrance cells? sub-cells with special flags?) — no symptom observed; audit it if a lighting edge case arises in an unusual cell type | `useSunlight` gate `DrawMeshInternal` 0x0059f398; `useSunlightSet` 0x0054d450; per-stage `PView::DrawCells` 0x005a4840 (`useSunlightSet(1)` 0x005a485a / `useSunlightSet(0)` 0x005a49f3); `minimize_object_lighting` 0x0054d480; `CellManager::ChangePosition` 0x004559B0 (ambient + seen_outside) |
 | AP-35 | Point/spot lights are now PER-VERTEX Gouraud (`pointContribution` ~line 153 of `mesh_modern.vert`) matching retail's `SetStaticLightingVertexColors` bake path. Half-Lambert wrap (`(1/1.5)·(N·D + 0.5·d)`) AND norm distance attenuation (`distsq>1 ? distsq·d : d`) ARE ported (A7 Fix A, `aa94ced`). Point-light sum clamped to [0,1] on its own accumulator before adding ambient+sun (A7 Fix D D-1, mirrors retail's per-vertex bake clamp). CPU oracle: `src/AcDream.Core/Lighting/LightBake.cs`, locked by `tests/AcDream.Core.Tests/Lighting/LightBakeConformanceTests.cs`. **Residual (two parts):** (a) acdream lights in-shader each frame (per-frame GPU evaluate); retail bakes into the vertex buffer ONCE — an architecture/performance difference; the wrap + norm + clamp formula is the same, but bake-once is cheaper for static geometry; (b) acdream's `SelectForObject` keeps only the 8 NEAREST reaching point/spot lights per object/cell (`MaxLightsPerObject=8`, see AP-16), whereas retail's bake sums ALL reaching static lights per vertex — a surface reached by >8 point lights is dimmer in acdream than retail's bake result (rare in practice; a room has a handful of torches) | `src/AcDream.App/Rendering/Shaders/mesh_modern.vert` (`pointContribution` ~line 153; wrap ~line 163; norm ~line 167; point-sum clamp line 210) | Per-vertex Gouraud + wrap + norm + clamp all match retail. The two residuals are: (a) per-frame GPU vs bake-once — architecture/perf only; (b) 8-light cap dimming when >8 lights reach one surface — rare. `LightInfoLoader.cs:81` folds static_light_factor 1.3 into Range | (a) A new frame-time consumer bypassing `accumulateLights` would need to replicate the wrap + norm formula; per-frame GPU re-evaluate has higher per-frame cost than bake for static geometry. (b) A densely lit scene (>8 torches reaching one wall) renders dimmer than retail — see AP-16 for the 8-cap ownership | `calc_point_light` 0x0059c8b0 (line 0x0059c9a2 ramp; 0x0059c925 wrap); `SetStaticLightingVertexColors` 0x0059cfe0; static_light_factor 0x00820e24 |
 | AP-37 | LayoutDesc meters collapse Type-3 slice descendants into `UiMeter.BackLeft..FrontRight` and reuse `UiMeter.DrawHBar` rather than building those media descendants and dispatching retail `UIElement_Meter::DrawChildren`. Non-Type-3 meter children are imported normally. | `src/AcDream.App/UI/Layout/DatWidgetFactory.cs` (`BuildMeter`/`SliceIds`); `LayoutImporter.cs` meter child predicate | The current vitals/character meter shapes are visually accepted and fixture-pinned; this is a representation adaptation, not a controller overlay. | A meter with a different descendant/media structure can render empty or with incorrect clipping/direction | `UIElement_Meter::DrawChildren @ 0x0046FBD0`; production meter LayoutDesc fixtures |