fix(render): complete S4 chunk 2 final alpha parity round

Final allowed fix round for S4-c2 on cc8e5677a. This lands every item in
the campaign packet section 12 without adding a flush site, shader, distance,
overflow recovery draw, or graphical-client run.

R2-1 particle row 5: ParticleRenderer now constructs and owns the actual
particle-mesh-opaque pipeline using the existing particle_mesh shaders and
layout, Blend=None, depth test/write enabled with WorldCompare, dynamic
per-batch cull, clockwise front face, and no alpha-to-coverage. The production
dispatch selects it for an opaque-classified mesh particle whose clamped
material alpha is 1.0. Nonopaque mesh pipelines remain depth-write-off. The
production route keeps cached reserve/immediate delegates and the warmed
append/immediate paths allocate 0 B.

R2-2 EnvCell exact per-subset routing: ObjectMeshManager carries Content's
TextureBatchData.RetailSurfaceMask onto ObjectRenderBatch at the real upload
boundary. EnvCellRenderer scans the active prepared cell snapshot and feeds
each real transparent batch's exact mask through RetailAlphaMeshRouter. Pure
0x08 Base1ClipMap reaches CLIP, 0x02 alpha-family reaches ALPHA, and table
Immediate subsets draw at the cell turn. Separate fixed-route draw sources
coalesce to at most one token per (cell,list) and filtered replay draws only
that list's subsets; a mixed cell contributes to both lists without duplicate
replay. Detail-on routes eligible subsets immediately with the detail pass.
The warmed dispatch/source allocation pins measure 0 B; the production scan/filter is covered behaviorally and uses only retained scratch/enumerators (static allocation audit).

R2-3 capacity cleanup: RetailAlphaQueue registers a source before the 3,000
entry capacity return. A source whose first append is rejected is therefore
reset by flush, EndFrame, or abort, but its rejected payload is never prepared
or drawn.

R2-4 production proof and prose: both actual Wb submit sites are exercised;
the particle tests call the production dispatcher and inspect the constructed
owner's production pipeline/selector; EnvCell tests upload real Content batch
masks through ObjectMeshManager and drain real filtered MDI calls. The A1
positive proof executes WorldSceneRenderer's real outdoor frame owner through
RetailPViewRenderer.DrawInside and RetailAlphaQueue.EndFrame and observes
[DrawBuilding x N, RenderNormalMode] with no LandscapeFlush. The packet and
register now state the varying retail first-for-list truth and the exact
per-subset EnvCell/AP-238, visible AP-239 compositing, and AP-240 feeder scope.
Physical active register counts remain AP=159 and AD=92.

Final clean-state gates (actual output):
- Release solution build: Build succeeded; 0 Warning(s); 0 Error(s).
- Hermetic solution filter: every project green, 16,728 passed / 0 failed /
  0 skipped total; AcDream.App.Tests 6,875/6,875.
- InstalledDat: 255 passed / 10 failed / 1 skipped / 266 total, exactly the
  allowed identities: TowerAscent_StaircaseStaysConeVisible_EveryStep;
  MainGameUiAndChatInput_MediaBearingChildrenNowBuildAsRealWidgets (#383);
  EveryAuthoredInvisibleWidget_StartsHiddenAcrossAllLayouts (#383);
  Oh_doorway_still_first_frame_diff (#458); and the six
  AlphaFlushCounts_{CathedralArrival,CathedralLeak,CathedralStairArch,
  FoundryDeep,HoltburgDoorwayStill,TerraceEdge}_MatchesRetailFrame2. All six
  AlphaFlushSites_* pass in the same lane.
- VulkanShaderDescriptorContractTests + VulkanShaderManifestTests +
  RenderPackSpirvValidatorTests: 32 passed / 0 failed / 0 skipped.
- Corrected queue/router/walk/driver/particle/Wb/EnvCell/PView production
  filter: 211 passed / 0 failed / 0 skipped.
- Explicit warmed production allocation pins: 2 passed; both measure 0 B.
- Register: physical AP-238/AP-239/AP-240/AD-120/~~AP-34~~ rows each count
  exactly 1; active physical rows AP=159 and AD=92.
- git diff --check: PASS.

Production mutation checks (each applied, run to the named first failure, and
exactly reversed before the final gates):

1. Restoring the particle row-5 throw fails
   OpaqueClassifiedMeshBatch_WithNoMaterialAlpha_DrawsImmediateOnOpaqueDepthState
   first with InvalidOperationException: mutation: row 5 unreachable.

2. Deleting row 5's immediate callback fails that same production-dispatch
   test's first collection assertion: expected [(Mesh, 11, True)], actual [].

3. Constructing the actual owner pipeline with depthWrite:false fails
   ImmediateOpaqueMesh_UsesProductionParticleMeshOpaquePipelineDescription
   first at Assert.True(description.Depth.Write): expected true, actual false.

4. Mapping the production selector back to _meshAlphaPipeline fails that same
   test first at Assert.Same: expected particle-mesh-opaque, actual
   particle-mesh-alpha.

5. Dropping RetailSurfaceMask at the real ObjectMeshManager upload boundary
   fails the mixed-cell production scan first: expected Clip | Alpha, actual
   Immediate.

6. Hardcoding the uploaded EnvCell scan to MaskAlphaFamily fails the mixed-cell
   production scan first: expected Clip | Alpha, actual Alpha; the pure-mask
   pin also reports expected Clip, actual Alpha.

7. Inverting the production detail predicate fails the detail-on production
   pin first: expected Immediate, actual Clip.

8. Removing both EnvCell replay filters fails the mixed production drain's real
   MDI assertions: each call expected DrawCount 1, actual DrawCount 2.

9. Moving RegisterSource below the full-capacity return fails every
   RejectedFirstUseSource_IsCleanedWithoutPrepareOrDraw row (flush, EndFrame,
   abort) at the first ResetCount assertion: expected 1, actual 0.

10. Hardcoding detailSurfaceActive=false at the real Wb dispatch fails both
    production submit-site tests at their first queue-count assertion:
    expected 0, actual 1.

11. Restoring RetailPViewRenderer's removed outdoor LandscapeFlush call fails
    OutdoorProductionPView_DrainsBuildingThenRenderNormalModeWithoutLandscapeFlush
    first at the real drain sequence: expected [DrawBuilding,
    RenderNormalMode], actual [DrawBuilding, LandscapeFlush,
    RenderNormalMode].

12. Deleting WorldSceneRenderer's final EndFrame owner drain fails that same
    full-path A1 test first: expected [DrawBuilding, RenderNormalMode], actual
    [DrawBuilding].

Not done/deferred: none. No retail conflict or infeasible contract item was
found. No graphical client was launched.

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

docs/architecture/retail-divergence-register.md

@@ -330,9 +330,9 @@ research and is no longer active.
 | ~~AP-28~~ | **RETIRED 2026-08-08 (Campaign A slice A2).** The three picked AL parameters and the gain-driven eviction are both gone. `RetailSoundMixer` now carries the byte-decoded retail curve — `g = dist < 5 ? vol : 25·vol/dist²`, clamped to 1, ONE master multiply, `db = ceil(20·log10 g)`, and a hard −50 dB no-allocate floor (audible radius ≈94.2 m at unity) — with pan as retail's `−15·sin(Δbearing)` in whole decibels and a 5-metre integer deadzone. Every AL source is source-relative with `AL_ROLLOFF_FACTOR = 0` and the global distance model is `None`, so AL contributes no attenuation of its own; the old `InverseDistanceClamped` ref-2 m curve was inverse FIRST power (`2/d`), quieter than retail up close and far louder at range with no cutoff at all. Voice eviction now compares the DAT-authored float priority strictly-less in ring order per `SoundManager::PlaySoundInternal` @ `0x0054FEC0` (the row's old `FUN_00550ad0` citation was wrong — that address is inside an `IntrusiveHashTable` constructor). The residual pan-LAW approximation is AP-173; retail's own `s_bPlaySoundOnlyWhenActive` gate is TS-64. | retired | — | — | `SoundManager::GetAttenuation @ 0x00550020`; `SoundManager::PlaySoundInternal @ 0x00550170` and `@ 0x0054FEC0`; `docs/research/2026-08-08-audio-retail-soundmanager-core.md` |
 | AP-31 | Scenery placement drift + the 0xA9B1 road-edge tree — WB-upstream divergences from retail, ACCEPTED (**#49/#50**, 2026-05-11) | `src/AcDream.Core/World/SceneryGenerator.cs` (via `WbSceneryAdapter`) | Piecemeal patching against WB upstream is net-negative (the `e279c46` road-check attempt over-suppressed scenery elsewhere, reverted `677a726`); visible impact = a handful of trees a few meters off | The same WB-upstream class could hide a *larger* placement divergence elsewhere; revisit only via a coherent ACME-style per-vertex filter port | `CLandBlock::get_land_scenes`; ACME GameScene.cs:1074 per-vertex road filter |
 | ~~AP-34~~ | **RETIRED 2026-09-04 (S4-c2 fix round 1) — residuals AP-238 / AP-239 / AP-240.** Historical: a single acdream-only distance-sorted alpha queue stood in for retail's two independent per-`DrawMesh`-invocation FIFO lists (CLIP/ALPHA, capacity 3000 each) drained at retail's four normal-world `FlushAlphaList` sites (`DrawBuilding`/`DrawBlock`/`PView::DrawCells`/`RenderNormalMode`). S4-c2 (`048d5b12f`) replaced it with `RetailAlphaQueue`'s exact two-list FIFO port of `D3DPolyRender::AddMeshToAlphaList`'s append-only/capacity-drop behavior, routed by `RetailAlphaMeshRouter`'s port of `DrawMesh`'s five-row branch table, under retail's exact no-op threshold rule — no acdream-only sort remains anywhere on the alpha path. | retired | — | — | `D3DPolyRender::AddMeshToAlphaList` 0x0059C230; `D3DPolyRender::FlushAlphaList` 0x0059D2E0; `D3DPolyRender::DrawMesh` 0x0059D4A0 |
-| AP-238 | **S4-c2 (2026-09-04); NARROWED FURTHER at the S4-c2 fix round 1 (2026-09-04) — renumbered from the colliding id "AP-236" (that id was already filed AND retired on `main` by an unrelated #132 fix; ids are never reused, per A2).** One of TWO residuals AP-34's closure left open (the other is AP-240's CLIP-list-for-ordinary-content gap) — NOT "the ONE residual" as originally claimed. `RetailAlphaQueue` is retail's own two independent FIFO lists (CLIP/ALPHA), routed by `RetailAlphaMeshRouter`'s port of `DrawMesh`'s branch table and drained at retail's four normal-world `FlushAlphaList` sites. Fix round 1 (M6) then wired EnvCell's transparent shell into that SAME table: `RetailPViewPassExecutor.SubmitOrDrawTransparentCellShell` routes through `RetailAlphaMeshRouter.Route` with the shell's mask (always 0x02, alpha-family) — detail ON takes Row 1 (immediate, unchanged from before); detail OFF now takes Row 3 (Append) instead of drawing immediately, appending ONE token per CELL via the new `EnvCellAlphaDrawSource`, replayed through `EnvCellRenderer.RenderTransparentOrdered` at the queue's next flush. The residual is now PURELY the token GRANULARITY, not "never through the shared queue" (that claim is retired) — no "substantial separate architecture change" was needed; `RenderTransparentOrdered` already replayed a cell's transparent batch on demand. | `src/AcDream.App/Rendering/RetailPViewPassExecutor.cs` (`SubmitOrDrawTransparentCellShell`, `EnvCellAlphaDrawSource`); `src/AcDream.App/Rendering/Wb/EnvCellRenderer.Rhi.cs` (`TransparentDetailEnabled`); `src/AcDream.App/Rendering/RetailAlphaQueue.cs`; `src/AcDream.App/Rendering/RetailAlphaMeshRouter.cs` | EnvCell's per-cell transparent batch is homogeneously mask 0x02 (EnvCell shells never carry a ClipMap bit in this path), so one atomic per-cell FIFO token is consistent with how WbDrawDispatcher/ParticleRenderer already submit at instance granularity, not per-D3D-surface-subset; a particle or ordinary GfxObj instance appended between two cell tokens keeps its own position (the queue's adjacent-only batching invariant) — only interleaving BETWEEN two subsets of the SAME cell's own batch is unavailable, and nothing needs that today. | A single EnvCell's transparent shell (glass, water, grates) with multiple distinct subsets composites those subsets in strict draw-call order RELATIVE TO EACH OTHER (no visual change from before this round); the coarser per-cell token only matters if a future feature needs to interleave other content between two subsets of that SAME cell's own batch. | `RenderDeviceD3D::DrawEnvCell` 0x0059F170; `D3DPolyRender::AddMeshToAlphaList` 0x0059C230; `D3DPolyRender::FlushAlphaList` 0x0059D2E0 |
-| AP-239 | **Filed 2026-09-04 (S4-c2); renumbered from the colliding id "AP-237" at the S4-c2 fix round 1 (2026-09-04, A2) — ids are never reused; the dangling "AP-236b" citation at `RetailAlphaMeshRouter.cs` (which never matched either filed id) is corrected to this one in the same fix.** `RetailAlphaMeshRouter.MaskFromTranslucencyKind` reconstructs retail's `ConstructMesh` queue-routing mask (spec: alpha-family bits always win; else ClipMap; else Translucent) from acdream's already-collapsed `TranslucencyKind`, which was built for a DIFFERENT priority chain (`TranslucencyKindExtensions.FromSurfaceType`'s blend-STATE selection: a "Translucent override" step checks Translucent+ClipMap-without-alpha-family FIRST and calls it `AlphaBlend`). A raw `Surface.Type` with Translucent(0x10) and Base1ClipMap(0x04) bits but no Alpha/InvAlpha/Additive bit therefore reconstructs to mask 0x02 (routes to ALPHA) here, while retail's real `ConstructMesh` mask is 0x08 (routes to CLIP) — `GroupKey` does not retain the raw ClipMap bit needed to distinguish the two `AlphaBlend` origins, and threading it through `GfxObjSubMesh`/`GroupKey`/every mesh-classification call site was judged out of this bounded chunk's scope. **S4-c2 fix round 1 (M7, 2026-09-04): the original cited example was DISPROVEN and replaced.** Cloud GfxObj `0x01004C35`'s surface `0x08000023` is `Type=0x10114` (ADDITIVE\|ALPHA\|TRANSLUCENT\|BASE1_CLIPMAP) — its alpha-family bits win in BOTH retail's real priority and this reconstruction, so it reproduces IDENTICALLY (mask 0x02 either way) and never demonstrated the gap. A 2026-09-04 scan of every Portal-dat `Surface` (6,152 scanned) found 27 REAL instances of Translucent+Base1ClipMap with no alpha-family bit (`Type=0x14`); cross-referencing the Portal-dat `GfxObj` table found surface `0x08000015` referenced by GfxObj `0x010001EC` — a real, confirmed instance, not the disproven cloud one. | `src/AcDream.App/Rendering/RetailAlphaMeshRouter.cs` (`MaskFromTranslucencyKind`'s own doc comment); `src/AcDream.Core/Meshing/TranslucencyKind.cs` (`FromSurfaceType`'s Translucent-override step) | Compositing-ORDER only (which FIFO list a subset enters), never a blend-mode or visual difference — the affected surface still renders alpha-blended exactly as retail's `SetSurface` configures it; only its position in the shared drain order can differ at a narrow overlap with other CLIP-list content. 27 of 6,152 scanned surfaces (2026-09-04) carry this shape — narrow, but not as narrow as "only the cloud." | A translucent+clipmap surface (e.g. GfxObj `0x010001EC`) interleaves with ordinary ALPHA-list content instead of CLIP-list content at a rare compositing overlap. | `D3DPolyRender::ConstructMesh` 0x0059DFA0 (mask priority); `D3DPolyRender::SetSurface` 0x0059C4D0 (the Translucent-override blend selection acdream's `TranslucencyKind` already matches) |
-| AP-240 | **Filed 2026-09-04 (S4-c2 fix round 1, M2).** `WbDrawDispatcher.IsOpaque(t) => t == Opaque \|\| t == ClipMap` (`WbDrawDispatcher.cs`) keeps every clip-mapped subset out of `_translucentDraws`/`SubmitWalkAlphaInstance`/`SubmitToAlphaQueue` entirely; retail's `ConstructMesh` gives clip-mapped surfaces mask 0x08 → `DrawMesh` row 3 → the CLIP list, and the S4-c2 capture evidence (six `*.alphadepth.log` files) shows 4,183 of 10,556 `AM` lines on CLIP (e.g. `(DrawBuilding,0,158,452)`: 158 real CLIP entries acdream's ordinary-content path reports as 0). Clip-mapped GfxObj/scenery/building subsets instead draw IMMEDIATELY on the `OpaqueAlphaToCoverage` pipeline (the pre-S4/VisualMaster-era translucency decision — not a pipeline change this round) and never enter retail's CLIP FIFO at all; the only CLIP-list feeders left after S4-c2 are clip-mapped MESH PARTICLES (M1's fix: `ParticleRenderer` applies no `IsOpaque`-style pre-filter, so a mesh particle classified `TranslucencyKind.ClipMap` genuinely reaches mask 0x08 and can append to CLIP). | `src/AcDream.App/Rendering/Wb/WbDrawDispatcher.cs` (`IsOpaque`); `src/AcDream.App/Rendering/RetailAlphaMeshRouter.cs` (`MaskClipMap` — the only reachable feeder is `ParticleRenderer`) | Depth-writing alpha-to-coverage cutouts composite correctly regardless of draw order for the overwhelming majority of clip-mapped content (doors, windows, foliage) — the divergence is compositing ORDER at a narrow overlap with genuine CLIP-list content (clip-mapped particles), not a missing/wrong cutout. Changing this is a VisualMaster-era pipeline decision, out of this bounded chunk's scope. | Clip-mapped GfxObj/scenery/building content composites in draw-call order relative to other opaque-pipeline content instead of interleaving through retail's deferred CLIP drain with clip-mapped particles/EnvCell content — edges differ only at a narrow overlap, never a missing cutout. | `D3DPolyRender::ConstructMesh` 0x0059DFA0 (mask priority, row 3 → CLIP); `D3DPolyRender::DrawMesh` 0x0059D4A0 (row 3); `D3DPolyRender::AddMeshToAlphaList` 0x0059C230 |
+| AP-238 | **S4-c2 (2026-09-04); NARROWED FURTHER at fix rounds 1 and 2 — renumbered from the colliding id "AP-236" (that id was already filed AND retired on `main`; ids are never reused).** `RetailAlphaQueue` is retail's own independent CLIP/ALPHA FIFO pair. Fix round 2 now retains each real EnvCell `ObjectRenderBatch`'s `TextureBatchData.RetailSurfaceMask` and routes every transparent subset independently through `RetailAlphaMeshRouter`: canonical `0x08` Base1ClipMap subsets enter CLIP, `0x02` alpha-family subsets enter ALPHA, table-Immediate subsets draw at the cell turn, and detail-active eligible subsets remain immediate with detail. A mixed cell can therefore contribute to both lists. The remaining residual is only token granularity: acdream coalesces to at most ONE token per `(cell,list)` and replays a list-filtered subset view, where retail appends one entry per surface subset. | `src/AcDream.App/Rendering/RetailPViewPassExecutor.cs` (`SubmitOrDrawTransparentCellShell`, `EnvCellAlphaDrawSource`); `src/AcDream.App/Rendering/Wb/EnvCellRenderer.cs` / `.Rhi.cs` (`GetTransparentRoutes`, `RouteTransparentBatch`, filtered `RenderTransparentOrdered`); `src/AcDream.App/Rendering/Wb/ObjectMeshManager.cs` (`ObjectRenderBatch.RetailSurfaceMask`); `src/AcDream.App/Rendering/RetailAlphaQueue.cs` | Per-list replay preserves the exact subset class and the queue's adjacent-only batching preserves interleaving around other sources; the approximation exists only within one cell/list when several same-list subsets are represented by one token. | At a genuine depth-overlap, collapsing multiple same-list subsets to one `(cell,list)` token can visibly change compositing if other translucent content should fall between those subsets in retail's FIFO. | `RenderDeviceD3D::DrawEnvCell` 0x0059F170; `D3DPolyRender::DrawMesh` 0x0059D4A0; `D3DPolyRender::AddMeshToAlphaList` 0x0059C230; canonical installed-DAT case `F4180104` surface `08000BFF` mask `0x08` |
+| AP-239 | **Filed 2026-09-04 (S4-c2); renumbered from the colliding id "AP-237" at fix round 1.** `RetailAlphaMeshRouter.MaskFromTranslucencyKind` reconstructs retail's `ConstructMesh` routing mask from acdream's already-collapsed `TranslucencyKind`. A raw `Surface.Type` with Translucent(0x10) and Base1ClipMap(0x04), but no Alpha/InvAlpha/Additive bit, collapses to `AlphaBlend` and reconstructs as mask `0x02` (ALPHA), while retail retains mask `0x08` (CLIP). `GroupKey` does not retain the raw ClipMap bit required to distinguish the two origins. The original cloud example was disproven; a 2026-09-04 scan found 27 real `Type=0x14` surfaces among 6,152, including surface `0x08000015` referenced by GfxObj `0x010001EC`. | `src/AcDream.App/Rendering/RetailAlphaMeshRouter.cs` (`MaskFromTranslucencyKind`); `src/AcDream.Core/Meshing/TranslucencyKind.cs` (`FromSurfaceType`) | The affected surface keeps the same blend state, but choosing the wrong FIFO changes compositing order; at a genuine overlap that order difference can be visibly different. | A translucent+clipmap surface (e.g. GfxObj `0x010001EC`) interleaves with ALPHA-list content instead of CLIP-list content and can composite differently at a depth overlap. | `D3DPolyRender::ConstructMesh` 0x0059DFA0 (mask priority); `D3DPolyRender::SetSurface` 0x0059C4D0 |
+| AP-240 | **Filed 2026-09-04 (S4-c2 fix round 1, M2); narrowed at fix round 2.** `WbDrawDispatcher.IsOpaque(t) => t == Opaque \|\| t == ClipMap` keeps clip-mapped GfxObj/scenery/building subsets out of `_translucentDraws`/`SubmitWalkAlphaInstance`/`SubmitToAlphaQueue`; retail gives those subsets mask `0x08` and appends them to CLIP. They instead draw immediately on acdream's `OpaqueAlphaToCoverage` pipeline. Fix round 2 proves CLIP has two other production feeder families: clip-mapped mesh particles and exact-mask EnvCell subsets (canonical `F4180104` / surface `08000BFF`). The residual is therefore only ordinary GfxObj/scenery/building content, not an empty or particle-only CLIP list. | `src/AcDream.App/Rendering/Wb/WbDrawDispatcher.cs` (`IsOpaque`); `src/AcDream.App/Rendering/Wb/EnvCellRenderer.cs` (`RouteTransparentBatch`); `src/AcDream.App/Rendering/ParticleRenderer.cs`; `src/AcDream.App/Rendering/RetailAlphaMeshRouter.cs` | The cutout itself remains present, but immediate opaque-path placement differs from retail's deferred FIFO order. Changing it is a VisualMaster-era pipeline decision beyond this bounded chunk. | At an overlap with deferred CLIP content from particles or EnvCells, clip-mapped GfxObj/scenery/building edges can visibly composite in the wrong relative order. | `D3DPolyRender::ConstructMesh` 0x0059DFA0 (mask priority, row 3 → CLIP); `D3DPolyRender::DrawMesh` 0x0059D4A0; `D3DPolyRender::AddMeshToAlphaList` 0x0059C230 |
 | 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) + `src/AcDream.App/World/LiveEntityHydrationPorts.cs` (`LiveEntityWorldOriginCoordinator.TryInitialize` — login pre-collapse) + `src/AcDream.App/Physics/LiveEntityNetworkUpdateController.cs` (`OnPosition` — first accepted canonical Position) + `GameWindow:AimTeleportDestination`/`IsSealedDungeonCell` (teleport pre-collapse and DAT predicate) + `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/WorldRenderFrameBuilder.cs` (`RuntimeWorldFrameEnvironmentPreparation.UpdateSunFromSky`) | 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 |