Merge remote-tracking branch 'origin/main'

# Conflicts:
#	README.md
#	manifest.json

.agents/handoffs/forge-windows-g2-4fde0e94.md

@@ -0,0 +1,31 @@
+# Forge (Windows) — rebuild engine main at 4fde0e94 (G2 pooled GPU cull + static shadow fixes)
+
+Mac `engine/macos-arm64` is engine `4fde0e94` (2026-10-18, `4.7.2.stable.mono.custom_build.4fde0e94d`). Windows is
+`9dcc5d458`. **The game pins `4.7.2@4fde0e94d`** (owner's call: publish Mac-only while Windows sleeps), and
+`shaders/tree_wind.gdshader` uses the new `STATIC_SHADOW_AT_REST` built-in, so a 9dcc5d45 engine both hits the
+mismatch dialog and fails that shader. Rebuild before anything else on Windows.
+
+## Build
+`tinqs pull` the engine and `git checkout 4fde0e94` (later main commits may be docs or G4 work; build the pinned hash).
+Flags from the `windows-x64` manifest block. Regenerate GodotSharp from that binary (`--generate-mono-glue
+modules/mono/glue`, `build_assemblies.py --godot-output-dir=bin --godot-platform=windows`) and publish the WHOLE
+`bin/GodotSharp` (`Api/` and `Tools/`).
+
+## New since 9dcc5d45 (not compiled on Windows yet)
+- G2 pooled GPU cull: `servers/rendering/gpu_cull_arena.h`, `renderer_rd/shaders/effects/multimesh_cull.glsl`,
+  `renderer_rd/effects/multimesh_cull.*`, `storage_rd/mesh_storage.*`, `forward_clustered/render_forward_clustered.cpp`.
+  Three shared storage arenas (each < 512 MiB, the D3D12 raw-buffer limit); no buffer device address.
+  Switch `rendering/occlusion_culling/gpu_cull_pooled` / `RenderingServer.multimesh_gpu_cull_set_pooled`, default off.
+- Fixed shadow LOD band for gpu-culled MultiMeshes (`gpu_cull_shadow_fixed_band`, `multimesh_set_gpu_shadow_band`), default off.
+- `STATIC_SHADOW_AT_REST` shader built-in, true only in static shadow tile passes when
+  `rendering/lights_and_shadows/directional_shadow/static_casters_at_rest` is on; default off.
+
+## Checks
+1. Oracle `bash tests/gpu_oracle/multimesh_cull_pool/run.sh <exe> d3d12` (Mac Metal: 4524/4524). This gates making
+   pool the default. Also re-run `tests/gpu_oracle/draw_indirect_count/run.sh <exe> d3d12` (expect 22/2 as before).
+2. Game in-world on the new pair: `tree_wind` compiles (no `Unknown identifier STATIC_SHADOW_AT_REST`).
+   `env gpucull status`; then the G0/G1 profile (`.agents/handoffs/2026-10-18-windows-g0-render-thread-profile.md`
+   in ariki-game) with states `"pool off=gpucull:pool off" "pool on=gpucull:pool on"`. Mac result: cull dispatches
+   258/422/464 → 2, fps +1/0/+1, render CPU −0.3 ms at the village; Windows records those dispatches on its render
+   thread, so expect a larger gain there. Report the table.
+3. Publish windows-x64 here (binary + console binary + complete GodotSharp, sha256s, manifest note, CRLF), push.

README.md

@@ -31,10 +31,10 @@ into GodotSharp, so GodotSharp had to be regenerated — and must ship here alon
 
 | Platform | Engine | Terrain3D | GodotSharp shipped? |
 |---|---|---|---|
-| `engine/macos-arm64/` | **4.7.2** (`9dcc5d45`) — screen-space contact shadows (directional), hero shadow min-combine, cascade caster-culling + shadow-bias fixes from upstream 4.8, `draw_list_draw_indirect_count` (Metal), on top of everything `7fb81364` had | built-in module | ✅ yes (`GodotSharp/`) |
-| `engine/windows-x64/`  | **4.7.2** (`4fde0e94d`) — everything `9dcc5d45` had **plus G2 pooled cull** (one reset + one cull dispatch per camera render, `gpu_cull_pooled` default off, fixed shadow band, STATIC_SHADOW_AT_REST); `template_release` here is still the `1e7be28b` tree | built-in module | ✅ yes (`GodotSharp/`) |
+| `engine/macos-arm64/` | **4.7.2** (`4fde0e94`) — G2 pooled GPU cull (one reset + one cull dispatch per camera render), fixed shadow LOD band and `STATIC_SHADOW_AT_REST` for static shadow tiles (all default off), on top of everything `9dcc5d45` had | built-in module | ✅ yes (`GodotSharp/`) |
+| `engine/windows-x64/`  | **4.7.2** (`4fde0e94d`) — same hash as macos-arm64: everything `9dcc5d45` had plus G2 pooled cull (`gpu_cull_pooled` default off), fixed shadow band, `STATIC_SHADOW_AT_REST` (tree_wind needs it); `template_release` here is still the `1e7be28b` tree | built-in module | ✅ yes (`GodotSharp/`) |
 
-Mac is engine `9dcc5d45` as of 2026-10-18 (`--version` prints `4.7.2.stable.mono.custom_build.9dcc5d458`): screen-space contact shadows for directional lights (default off; game dial `env contactshadow`), hero shadow min-combine, upstream 4.8 shadow fixes #121775 (cascade caster culling) and #124348 (blur < 1 no longer shrinks the bias; the game sun bias is 0.014 with it), and `draw_list_draw_indirect_count` with Metal emulation. GodotSharp regenerated (API + Tools). **Windows is one commit ahead: `4fde0e94d` (G2 pooled cull)** as of 2026-10-18 — Forge built engine main first; the game pin is version-only `4.7.2` until the Mac publishes `4fde0e94d` (then pin `4.7.2@4fde0e94d`). Windows oracle notes: `draw_indirect_count` on D3D12 at `9dcc5d458` was 22 passed / 2 failed (base-instance pair, the documented D3D12 gap, `.agents/handoffs/forge-windows-sscs-9dcc5d45.md`); `--gpu-validation` needs the D3D12 debug layer, which Forge lacks (Graphics Tools not installed); `multimesh_cull_pool` oracle not yet run on Windows. In-world checks (`herocombine`, `contactshadow`, SSCS in `--gpu-profile`, sun acne at bias 0.014, pooled gpucull census) still owed on Windows.
+Mac is engine `4fde0e94` as of 2026-10-18 (`--version` prints `4.7.2.stable.mono.custom_build.4fde0e94d`): G2 pooled GPU cull over shared storage arenas (`rendering/occlusion_culling/gpu_cull_pooled`, game dial `env gpucull pool`), a fixed shadow LOD band for gpu-culled MultiMeshes (`gpu_cull_shadow_fixed_band`, `env gpucull shadowband`) and the `STATIC_SHADOW_AT_REST` shader built-in (`static_casters_at_rest`, `env gpucull staticrest`); all default off. GodotSharp regenerated (API + Tools). Oracles on Metal: `multimesh_cull_pool` 4524/4524, `draw_indirect_count` 24/24. Windows is also engine `4fde0e94` as of 2026-10-18 — Forge built it in parallel with the Mac (`.agents/handoffs/forge-windows-g2-4fde0e94.md`), so both platforms and the `4.7.2@4fde0e94d` pin (the game's `tree_wind.gdshader` uses the new built-in) share one hash. Underneath is the `9dcc5d458` pair (SSCS contact shadows, hero shadow min-combine, upstream 4.8 shadow fixes #121775/#124348 with the game sun bias 0.014, `draw_list_draw_indirect_count`), where the D3D12 oracle on RTX 5080 was 22 passed / 2 failed — the base-instance pair, the documented D3D12 gap (`.agents/handoffs/forge-windows-sscs-9dcc5d45.md`); `--gpu-validation` needs the D3D12 debug layer, which Forge lacks (Graphics Tools not installed). Still owed on Windows: `multimesh_cull_pool` oracle (gates making pool the default), the validation pass, and in-world checks (`herocombine`, `contactshadow`, SSCS in `--gpu-profile`, sun acne at bias 0.014, pooled gpucull census).
 
 Long-term: CI (Gitea Actions) builds on push and publishes here / to S3 —
 `docs/ai/plans/build-pipeline-architecture-2026-06-07.md`.

manifest.json

@@ -4,16 +4,16 @@
     "macos-arm64": {
       "binary": "engine/macos-arm64/tinqs.macos.editor.arm64.mono",
       "godotsharp": "engine/macos-arm64/GodotSharp",
-      "version": "4.7.2.stable.mono.custom_build.9dcc5d458",
+      "version": "4.7.2.stable.mono.custom_build.4fde0e94d",
       "engine_repo": "tinqs/engine",
-      "engine_commit": "9dcc5d45",
-      "sha256": "2da4c3a6dc363375033bc20fdb1604d5a0de896d580da76e39305c6aade2ef7f",
+      "engine_commit": "4fde0e94",
+      "sha256": "a37ceb6d8aa5772f50c61cad12ee738881b9d060df00d9f392ae40b2437d8e54",
       "target": "editor",
       "built_on": "2026-10-18",
-      "built_by": "Ozan/Kraken via Claude (SSCS + hero shadow combine + shadow fixes + draw_indirect_count)",
+      "built_by": "Ozan/Kraken via Claude (G2 pooled GPU cull + static shadow fixed band / casters at rest)",
       "toolchain": "Xcode (Apple Silicon), SCons, VulkanSDK 1.4.350.0",
       "scons_flags": "platform=macos target=editor arch=arm64 module_mono_enabled=yes vulkan_sdk_path=~/VulkanSDK/1.4.350.0/macOS",
-      "note": "Mac. Engine 9dcc5d45 = 7fb81364 + screen-space contact shadows for directional lights (backport of upstream 4.8 #118045/#123814; ProjectSettings rendering/lights_and_shadows/contact_shadow/{enabled,shadow_length,surface_thickness}, Light3D.shadow_contact_shadows_allow/blur, RenderingServer.light_set_allow_contact_shadows; default off, game dial env contactshadow; ~0.43 ms on M1 Max), hero shadow min-combine (RenderingServer.instance_light_directional_set_hero_shadow_spot), tight cascade caster culling on the real cascade frustum (#121775), shadow bias no longer shrunk by blur < 1 (#124348; the game retunes the sun ShadowBias 0.04 -> 0.014 with the pin), RenderingDevice.draw_list_draw_indirect_count (Metal emulation; oracle 24/24 on M1 Max). GodotSharp regenerated from this binary (API + Tools). Smoke: the owner's game in-world, horizon + static prefill OK, hero combine on, contact shadows on/off, ~72 fps, no new error types. windows-x64 is STILL 7fb81364b: Forge must rebuild from engine main (handoff .agents/handoffs/forge-windows-sscs-9dcc5d45.md). The game pin is 4.7.2@9dcc5d458, so Windows engines are refused until that publish (owner's call, 2026-10-18)."
+      "note": "Mac. Engine 4fde0e94 = 9dcc5d45 + G2 pooled GPU cull (one reset + one cull dispatch per camera render over shared arenas; rendering/occlusion_culling/gpu_cull_pooled, default off) + fixed shadow LOD band for gpu-culled MultiMeshes (gpu_cull_shadow_fixed_band, default off) + STATIC_SHADOW_AT_REST shader built-in for static shadow tiles (rendering/lights_and_shadows/directional_shadow/static_casters_at_rest, default off). GodotSharp regenerated (with Tools/). Oracles: multimesh_cull_pool 4524/4524, draw_indirect_count 24/24 on Metal. Windows at 9dcc5d45; Forge to rebuild 4fde0e94 (handoff forge-windows-g2-4fde0e94.md)."
     },
     "windows-x64": {
       "binary": "engine/windows-x64/tinqs.windows.editor.x86_64.mono.exe",
@@ -26,10 +26,10 @@
       "sha256_console": "192aba5d6028fbf781fc0c41ff1cfc4bd5b51a3d84f99262b07bf6d61d9981ce",
       "target": "editor",
       "built_on": "2026-10-18",
-      "built_by": "Ozan/Forge via ZCode (G2 pooled cull, first publisher — no Forge handoff written for this one)",
+      "built_by": "Ozan/Forge via ZCode (forge-windows-g2-4fde0e94 handoff; built in parallel with the Mac)",
       "toolchain": "MSVC 14.3 (VS2022 Build Tools), Windows SDK 10.0.26100, SCons 4.10.1, .NET SDK 10.0.103",
       "scons_flags": "platform=windows target=editor module_mono_enabled=yes -j32",
-      "note": "Windows carries 4fde0e94d — engine main = 9dcc5d45 (SSCS + hero combine + shadow fixes + draw_indirect_count, see that pair) PLUS 4fde0e94d7 G2 pooled cull: one reset + one cull dispatch per camera render over shared arenas (multimesh_gpu_cull_set_pooled / project setting rendering/occlusion_culling/gpu_cull_pooled, DEFAULT OFF), fixed shadow band (multimesh_gpu_cull_set_shadow_fixed_band, multimesh_set_gpu_shadow_band), STATIC_SHADOW_AT_REST spatial built-in, multimesh_get_cull_command_offset; either shadow switch re-renders every static tile. Validated on M1 Max Metal pre-push (doctests 7/7, multimesh_cull_pool oracle 4524/4524, draw_indirect_count 24/24, StaticShadow 21/21). GodotSharp regenerated from this binary — Api AND Tools (MultimeshGpuCullSetPooled + MultimeshGpuCullSetShadowFixedBand + MultimeshSetGpuShadowBand present). Headless game boot clean (0 binding errors). Windows oracle for multimesh_cull_pool NOT yet run (D3D12 base-instance gap may bite the pooled draws — owed). Game pin set to version-only 4.7.2: macos-arm64 is still 9dcc5d45 — when Kraken publishes 4fde0e94d, pin 4.7.2@4fde0e94d. In-world gpucull checks (pooled on/off, census via multimesh_get_cull_command_offset, shadow band) still owed on Windows. Editor target only — template_release still 1e7be28b. D3D12 Agility DLLs not shipped (Vulkan game). Crash-report checklist (8 steps) also still owed."
+      "note": "Windows carries 4fde0e94d — SAME hash as macos-arm64 (built at 4fde0e94d7): engine main = 9dcc5d45 (SSCS + hero combine + shadow fixes + draw_indirect_count — for the 9dcc5d458 pair's Windows checks see its note in git history; oracle draw_indirect_count d3d12 was 22 passed / 2 failed, the base-instance pair being the documented D3D12 gap) PLUS 4fde0e94d7 G2 pooled cull: one reset + one cull dispatch per camera render over shared arenas (multimesh_gpu_cull_set_pooled / project setting rendering/occlusion_culling/gpu_cull_pooled, DEFAULT OFF), fixed shadow band (multimesh_gpu_cull_set_shadow_fixed_band, multimesh_set_gpu_shadow_band), STATIC_SHADOW_AT_REST spatial built-in (tree_wind.gdshader needs it), multimesh_get_cull_command_offset; either shadow switch re-renders every static tile. Validated on M1 Max Metal pre-push (doctests 7/7, multimesh_cull_pool oracle 4524/4524, draw_indirect_count 24/24, StaticShadow 21/21). GodotSharp regenerated from this binary — Api AND Tools (MultimeshGpuCullSetPooled + MultimeshGpuCullSetShadowFixedBand + MultimeshSetGpuShadowBand present). Headless game boot clean (0 binding errors). Owed on Windows: multimesh_cull_pool oracle (D3D12 base-instance gap may bite the pooled draws; gates making pool the default), --gpu-validation pass (D3D12 debug layer / Graphics Tools not installed on Forge), in-world gpucull census (multimesh_get_cull_command_offset) + pooled on/off, tree_wind compile check. Editor target only — template_release still 1e7be28b. D3D12 Agility DLLs not shipped (Vulkan game). Crash-report checklist (8 steps) also still owed."
     }
   }
-}
+}