streaming: deeper ready queue + startup apply cap; tile-local mouth-support culling (L72)

- Unreal: ready queue depth 8 -> 16 (cap 4 worker-sets / 64), startup mesh-apply cap
  12 -> 24 within the unchanged 14 ms budget. Owner SW arrival, same 2,850 keys:
  initial set ready 4.78 -> 3.74 s, queue p95 123 -> 35 ms, request-to-apply p95
  139 -> 45 ms; walkable gate 1.07 -> 1.11 s; per-frame apply p95 0.73 ms.
- Core composer: mouth-support segments filtered by tile bounds, mouth relief result reused
  (route checks -80/-87% in passage tiles). Hashes identical W1/W8/W12, Clang and GCC.
- Opt-in per-tile composer counters. Verify control meshes stored as .xz (78.5 -> 18.4 MB).
- Fast tier 269 PASS.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016dhxM43dnWvUu2B5Ph1iG1

Standalone/Source/StreamingWorld.h

@@ -710,6 +710,8 @@ public:
     {bProfileMesherStages=Enabled;}
     void SetTileDrainInstrumentationEnabled(bool Enabled)
     {bInstrumentTileDrain=Enabled;}
+    void SetPassageComposerProfilingEnabled(bool Enabled)
+    {bProfilePassageComposer=Enabled;}
     void SetExactMidpointDensityMemoEnabled(bool Enabled)
     {bExactMidpointDensityMemo=Enabled;}
     void SetActiveLaneBatchingEnabled(bool Enabled)
@@ -890,6 +892,7 @@ private:
             Options.useSharedColumnCache=bSharedColumnCache;
             Options.lod8HorizonInteriorStepVoxels=Lod8HorizonInteriorStepVoxels;
             Options.profileMesherStages=(bProfileMesherStages||bInstrumentTileDrain)&&!ForceGenerate;
+            Options.profilePassageComposer=bProfilePassageComposer;
             Options.instrumentTileDrain=bInstrumentTileDrain&&!ForceGenerate;
             Options.exactMidpointDensityMemo=bExactMidpointDensityMemo&&!ForceGenerate;
             Options.activeLaneBatching=bActiveLaneBatching&&!ForceGenerate;
@@ -1100,6 +1103,34 @@ private:
             static_cast<unsigned long long>(D.islandSharedCacheComputes),
             static_cast<unsigned long long>(D.islandSharedCacheConflictReplacements),
             static_cast<double>(D.islandSharedCacheLockWaitNanoseconds)/1000.0);
+        if(D.passageComposerProfiled)
+            std::printf("[PassageComposerTile] key=(%d,%d,%d,lod%d) worker_ms=%.3f samples=%llu native_field_calls=%llu features=%llu route_candidates=%llu mouth_candidates=%llu queries=%llu passage_iterations=%llu upper_mouth_checks=%llu column_candidates=%llu column_distance_calls=%llu negative_bounds=%llu negative_rejects=%llu positive_bounds=%llu positive_rejects=%llu carve_segment_visits=%llu smooth_carve_joins=%llu support_segment_visits=%llu smooth_support_joins=%llu support_capsule_checks=%llu later_bore_checks=%llu other_capsule_checks=%llu mouth_support_calls=%llu mouth_route_segment_checks=%llu relief_field_calls=%llu smoothstep_evaluations=%llu\n",
+                Event.Key.Coord.X,Event.Key.Coord.Y,Event.Key.Coord.Z,Event.Key.Level,
+                Event.WorkerMilliseconds,static_cast<unsigned long long>(Event.DensitySamples),
+                static_cast<unsigned long long>(D.fieldCalls),
+                static_cast<unsigned long long>(D.passageTileFeatureCount),
+                static_cast<unsigned long long>(D.passageTileRouteSegmentCandidates),
+                static_cast<unsigned long long>(D.passageTileMouthSegmentCandidates),
+                static_cast<unsigned long long>(D.passageComposerQueries),
+                static_cast<unsigned long long>(D.passageIterations),
+                static_cast<unsigned long long>(D.passageUpperMouthFilterChecks),
+                static_cast<unsigned long long>(D.passageColumnCandidates),
+                static_cast<unsigned long long>(D.passageColumnDistanceCalls),
+                static_cast<unsigned long long>(D.passageNegativeBoundChecks),
+                static_cast<unsigned long long>(D.passageNegativeBoundRejects),
+                static_cast<unsigned long long>(D.passagePositiveBoundChecks),
+                static_cast<unsigned long long>(D.passagePositiveBoundRejects),
+                static_cast<unsigned long long>(D.passageCarveSegmentVisits),
+                static_cast<unsigned long long>(D.passageSmoothCarveJoins),
+                static_cast<unsigned long long>(D.passageSupportSegmentVisits),
+                static_cast<unsigned long long>(D.passageSmoothSupportJoins),
+                static_cast<unsigned long long>(D.passageSupportCapsuleChecks),
+                static_cast<unsigned long long>(D.passageLaterBoreChecks),
+                static_cast<unsigned long long>(D.passageOtherCapsuleChecks),
+                static_cast<unsigned long long>(D.passageMouthSupportCalls),
+                static_cast<unsigned long long>(D.passageMouthRouteSegmentChecks),
+                static_cast<unsigned long long>(D.passageReliefFieldCalls),
+                static_cast<unsigned long long>(D.passageSmoothstepEvaluations));
     }
     void RecordTimelineEvent(FStandaloneCpuTileTimelineEvent::EType Type,
         const FVoxelTileKey& Key,int32 WorkerId,size_t QueueDepth,size_t ActiveWorkers,
@@ -1138,6 +1169,7 @@ private:
     int32 Lod8HorizonInteriorStepVoxels=0;
     bool bProfileMesherStages=false;
     bool bInstrumentTileDrain=false;
+    bool bProfilePassageComposer=false;
     bool bExactMidpointDensityMemo=false;
     bool bActiveLaneBatching=false;
     int32 ProofBlockCells=8;