fix(render): S3 chunk 3 re-review follow-ups

- the LandCell event doc states the deferred cross-block batch rule (the
  same-landblock lookahead text was stale);
- the AlphaSubmitMark arm's comment gives the true reason it is not a
  flush point (it only enqueues into the CPU alpha list; the drain leaves
  flush first);
- the pending terrain batch is cleared with the other transient frame
  lists in AbortFrame/BeginFrame;
- CompleteWalkTerrainFrame runs in a finally so a throwing Replay cannot
  leak its accumulated ticks into the next frame's sample;
- CopyRenderableEmittersInCell keeps its own doc comment;
- the outdoor-root LandCell pin now observes one real cell turn after its
  terrain;
- the oh1 landscape contract's stale "whole pre-stage" prose is resolved.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

docs/research/2026-09-01-overhaul/oh1-construction-landscape-contract.md

@@ -311,10 +311,11 @@ weather/sky post-pass
 | Cell object draw consumes the cell's exact stable far-to-near `shadow_part_list`. | The driver materializes per-cell record buckets, classifies/batches them, and uses a scope-level ordered stream plus separate alpha collection. | OH3/OH7 must replace reconstructed membership/order with exact part entries and exact per-cell alpha barriers. |
 | Building is alpha drain -> portal pass -> shell, then cell objects. | `RetailFrameWalk.DrawBuilding` and the driver's building events model this order. | Mechanism broadly matches, but its correctness depends on the sibling OH1 built-mesh/view and depth-lifecycle findings. |
 
-The architecture SSOT currently says that the walk records one retail-ordered
-stream including landscape cells, but the implementation and its own comments
-admit that terrain is one whole pre-stage. That is an architecture/code
-contradiction, not a documented adaptation. The same architecture section
+The architecture SSOT says that the walk records one retail-ordered stream
+including landscape cells; until S3 chunk 3 the implementation drew terrain as
+one whole pre-stage (an architecture/code contradiction, RESOLVED 2026-09-03:
+`RetailFrameWalk.DrawLandscape` now emits one land-cell terrain turn per
+admitted cell before that cell's own building/object turn). The same architecture section
 also states that every admitted portal slice GPU-clips the mesh; the dirty
 built-mesh candidate says the opposite. The sibling built-mesh evidence must
 arbitrate that claim before the architecture text or production path is

src/AcDream.App/Rendering/RetailPViewRenderer.cs

@@ -588,14 +588,21 @@ internal sealed class RetailPViewRenderer
         // filter) is gone — the walk's own per-cell LandCell turns are now
         // the sole terrain-visibility authority, so there is no plumbing
         // left to feed here.
-        driver.Replay(frame, encoder);
-        // S3 chunk 3 fix round 1 (F3): "the end of the walk replay" —
-        // exactly where the deleted whole-stage terrain leaf's own
-        // Begin()/Complete() bracket used to close. Pushes this frame's ONE
-        // accumulated terrain-timing sample (every DrawWalkLandCellBatch
-        // call between here and the last CompleteWalkTerrainFrame) and
-        // publishes the periodic [TERRAIN-DIAG] line if the cadence is due.
-        passes.CompleteWalkTerrainFrame();
+        try
+        {
+            driver.Replay(frame, encoder);
+        }
+        finally
+        {
+            // S3 chunk 3 fix round 1 (F3): "the end of the walk replay" —
+            // exactly where the deleted whole-stage terrain leaf's own
+            // Begin()/Complete() bracket used to close. Pushes this frame's
+            // ONE accumulated terrain-timing sample (every DrawWalkLandCellBatch
+            // call since the last CompleteWalkTerrainFrame) and publishes the
+            // periodic [TERRAIN-DIAG] line if the cadence is due. In a finally
+            // so a throwing Replay cannot leak its ticks into the next frame.
+            passes.CompleteWalkTerrainFrame();
+        }
 
         // Landscape-stage static-owner particles (candles, the cathedral
         // falls) submit AT THEIR OWN WALK TURNS inside Replay

src/AcDream.App/Rendering/Walk/WalkFrameDriver.cs

@@ -282,15 +282,16 @@ internal enum WalkFrameEventKind : byte
     /// <summary><see cref="IWalkFrameLeafRenderer.DrawSky"/>.</summary>
     Sky,
 
-    /// <summary>S3 chunk 3: <see cref="IWalkFrameLeafRenderer.DrawLandCellBatch"/>
-    /// for ONE admitted land cell — <see cref="WalkFrameEvent.CellId"/> is
-    /// the owning landblock id, <see cref="WalkFrameEvent.IntArg"/> packs
-    /// <c>(sideCellCount &lt;&lt; 8) | cellIndex</c> (side ∈ {1,2,4,8},
-    /// cellIndex &lt; side², both fit comfortably below the 8-bit shift).
-    /// <see cref="Replay"/> merges a run of consecutive same-landblock
-    /// entries into ONE <see cref="IWalkFrameLeafRenderer.DrawLandCellBatch"/>
-    /// call (B2's order-preserving batching) rather than replaying them
-    /// one at a time.</summary>
+    /// <summary><see cref="IWalkFrameLeafRenderer.DrawLandCellBatch"/> —
+    /// one admitted land cell's terrain (retail <c>DrawLandCell</c>
+    /// @0x0059f120): <see cref="WalkFrameEvent.CellId"/> is the landblock,
+    /// <see cref="WalkFrameEvent.IntArg"/> packs <c>side &lt;&lt; 8 | cellIndex</c>.
+    /// <see cref="Replay"/> does NOT draw it immediately: it appends the
+    /// cell to ONE pending terrain batch that may span landblocks and flushes
+    /// that batch only before a leaf that will submit GPU work (S3 chunk 3
+    /// fix round 1, §9.6 F2) — a particle turn whose cell has no renderable
+    /// emitter submits nothing and keeps the batch open. The GPU submission
+    /// order is therefore identical to replaying each cell alone.</summary>
     LandCell,
 
     /// <summary><see cref="IWalkFrameLeafRenderer.DrawCellShell"/> —
@@ -741,6 +742,7 @@ internal sealed class WalkFrameDriver : IWalkEventSink, IWalkLookInViewSource
         _events.Clear();
         _markPositions.Clear();
         _alphaSubmissions.Clear();
+        _pendingTerrainBatch.Clear();
         _alphaSubmitMark = 0;
         VisitedCells.Clear();
         LookInCellTurns.Clear();
@@ -877,6 +879,7 @@ internal sealed class WalkFrameDriver : IWalkEventSink, IWalkLookInViewSource
         _events.Clear();
         _markPositions.Clear();
         _alphaSubmissions.Clear();
+        _pendingTerrainBatch.Clear();
         _alphaSubmitMark = 0;
         VisitedCells.Clear();
         LookInCellTurns.Clear();
@@ -968,12 +971,14 @@ internal sealed class WalkFrameDriver : IWalkEventSink, IWalkLookInViewSource
                         cursor = end;
                         break;
                     case WalkFrameEventKind.AlphaSubmitMark:
-                        // Not a flush point (F2): an AlphaSubmitMark event is
-                        // always recorded alongside — immediately after — a
-                        // StreamMark from the SAME MarkIfGrown/MarkAlphaIfGrown
-                        // pairing (every Collect-side call site pairs them),
-                        // so the StreamMark case above already flushed any
-                        // pending batch by the time this one runs.
+                        // Not a flush point (F2): this arm only ENQUEUES the
+                        // translucent instances into the CPU alpha list
+                        // (SubmitWalkAlphaInstance); nothing reaches the GPU
+                        // until the list drains at an AlphaBarrier /
+                        // LandscapeFlush leaf or the pass end — each of which
+                        // flushes the pending terrain batch first, so the
+                        // drained alpha still lands after every preceding
+                        // land cell's terrain, exactly as when unbatched.
                         int alphaEnd = e.IntArg;
                         for (; alphaCursor < alphaEnd; alphaCursor++)
                         {
@@ -1064,6 +1069,7 @@ internal sealed class WalkFrameDriver : IWalkEventSink, IWalkLookInViewSource
             _events.Clear();
             _markPositions.Clear();
             _alphaSubmissions.Clear();
+        _pendingTerrainBatch.Clear();
             _alphaSubmitMark = 0;
             _readyToReplay = false;
             _ctx = null;

src/AcDream.Core/Vfx/ParticleSystem.cs

@@ -606,17 +606,6 @@ public sealed class ParticleSystem : IParticleSystem
         destination.Sort(static (left, right) => left.Handle.CompareTo(right.Handle));
     }
 
-    /// <summary>
-    /// Copies the renderable emitters whose <see cref="ParticleEmitter.OwnerCellId"/>
-    /// is <paramref name="cellId"/>, retaining spawn order. Retail
-    /// <c>CPhysicsObj::add_particle_shadow_to_cell</c> (0x00514a70) gives an
-    /// emitter exactly one shadow in its own current cell, independent of its
-    /// attached owner's registry membership — a hidden/suspended owner's
-    /// emitter is still enumerated here as long as it remains renderable
-    /// (presentation-visible and view-eligible). No per-call allocation after
-    /// warmup: the bucket's own sorted handle list is copied through the
-    /// retained <see cref="_scopeHandleScratch"/> buffer.
-    /// </summary>
     /// <summary>
     /// S3 chunk 3 fix round 1 (F2): whether <paramref name="cellId"/> has any
     /// renderable emitter for <paramref name="renderPass"/>, WITHOUT copying
@@ -635,6 +624,17 @@ public sealed class ParticleSystem : IParticleSystem
             && bucket.FirstRenderableHandle != 0;
     }
 
+    /// <summary>
+    /// Copies the renderable emitters whose <see cref="ParticleEmitter.OwnerCellId"/>
+    /// is <paramref name="cellId"/>, retaining spawn order. Retail
+    /// <c>CPhysicsObj::add_particle_shadow_to_cell</c> (0x00514a70) gives an
+    /// emitter exactly one shadow in its own current cell, independent of its
+    /// attached owner's registry membership — a hidden/suspended owner's
+    /// emitter is still enumerated here as long as it remains renderable
+    /// (presentation-visible and view-eligible). No per-call allocation after
+    /// warmup: the bucket's own sorted handle list is copied through the
+    /// retained <see cref="_scopeHandleScratch"/> buffer.
+    /// </summary>
     public void CopyRenderableEmittersInCell(
         ParticleRenderPass renderPass,
         uint cellId,

tests/AcDream.App.Tests/Rendering/Walk/WalkFrameDriverTests.cs

@@ -1360,7 +1360,12 @@ public sealed class WalkFrameDriverTests
         using var fx = new DispatcherFixture();
         var log = new List<string>();
         var leaf = new RecordingLeafRenderer(log);
+        // Every covered bucket but ONE reports no renderable emitter, so the
+        // block's object turns submit nothing (no flush) except bucket
+        // 0xF4180001, whose particle turn is a real submission — it must land
+        // AFTER the coarse cell's own terrain (retail DrawLandCell → DrawSortCell).
         leaf.CellsWithoutEmitters.UnionWith(CoarseLandscapeBuckets(0xF4180000u));
+        leaf.CellsWithoutEmitters.Remove(0xF4180001u);
         var ctx = new TestContext();
         var driver = new WalkFrameDriver(fx.Dispatcher, leaf, new FakeWorldData());
         var walk = new RetailFrameWalk();
@@ -1378,7 +1383,7 @@ public sealed class WalkFrameDriverTests
         driver.EndFrame();
         driver.Replay(draw.Frame, draw.Pass);
 
-        Assert.Equal(new[] { "SKY", "LANDCELL:f4180000:1:0" }, log);
+        Assert.Equal(new[] { "SKY", "LANDCELL:f4180000:1:0", "PARTICLES:f4180001" }, log);
         Assert.Equal(new (uint LandblockId, int SideCellCount, int CellIndex)[] { (0xF4180000u, 1, 0) },
             Assert.Single(leaf.LandCellBatches));
     }