acdream/src/AcDream.App/Rendering/TerrainModernRenderer.cs

using System.Numerics;
using AcDream.App.Rendering.Gpu;
using AcDream.App.Rendering.Wb;
using AcDream.Core.Terrain;

namespace AcDream.App.Rendering;

/// <summary>
/// Phase N.5b modern terrain dispatcher. Single global vertex/index arena with
/// a slot allocator (one slot per landblock, 384 verts × 40 bytes = 15,360
/// bytes per slot). Per-frame: build a DrawElementsIndirectCommand array from
/// visible slots and dispatch via one multi-draw-indirect call. Atlas
/// textures bound via the device's global texture table.
///
/// <para>Campaign V slice V11 deleted the raw-GL submission arm
/// (<c>TerrainModernRenderer.cs</c>'s former GL fields/constructor/Draw/Dispose
/// bodies); the RHI arm this file now exclusively hosts the shared allocator
/// and visibility logic for is defined in <c>TerrainModernRenderer.Rhi.cs</c>.</para>
/// </summary>
public sealed partial class TerrainModernRenderer : IDisposable
{
    // VertsPerLandblock MUST stay divisible by 6 — terrain_modern.vert uses
    // `gl_VertexID % 6` to pick the cell-corner index (BL/BR/TR/TL), and
    // because we bake `slot * VertsPerLandblock` into indices CPU-side and
    // pass BaseVertex=0 to MultiDrawElementsIndirect, gl_VertexID becomes
    // `slot * VertsPerLandblock + local_index`. The shader's modulo-6 only
    // reduces to `local_index % 6` because 384 is a multiple of 6. Changing
    // either constant without auditing the shader will silently mis-render.
    private const int VertsPerLandblock = LandblockMesh.VerticesPerLandblock;  // 384 (= 64 cells * 6 verts)
    private const int IndicesPerLandblock = VertsPerLandblock;
    private const int VertexSize = 40;  // sizeof(TerrainVertex)
    private const int IndexSize = sizeof(uint);
    private const float LandblockSize = LandblockMesh.LandblockSize;  // 192

    private readonly TerrainAtlas _atlas;

    /// <summary>A.5 T22.5: exposes the terrain atlas so callers can update
    /// anisotropic level mid-session via <see cref="TerrainAtlas.SetAnisotropic"/>.</summary>
    public TerrainAtlas Atlas => _atlas;

    private readonly GpuRetiredTerrainSlotAllocator _alloc;
    private readonly GpuRetirementLedger _retirementLedger;
    private bool _disposed;

    // Per-slot live data (index by slot integer; null entries are unused slots).
    private SlotData?[] _slots;

    // Reverse map: landblockId -> slot, for RemoveLandblock and replacement.
    private readonly Dictionary<uint, int> _idToSlot = new();

    // Backing-store capacity bookkeeping, shared with the RHI arm's arena.
    private long _globalVboCapacityBytes;
    private long _globalEboCapacityBytes;

    // Per-GPU-fenced-frame-slot draw bookkeeping, shared with the RHI arm.
    private int _dynamicFrameSlot;
    private bool _dynamicFrameStarted;

    /// <summary>
    /// The dynamic per-frame-slot indirect-command buffer pool this used to
    /// report was raw-GL-only bookkeeping, deleted with that arm at Campaign V
    /// slice V11. The RHI arm allocates its indirect-command storage from the
    /// GPU frame's own upload ring instead, so there is no separate pool to
    /// count.
    /// </summary>
    internal int DynamicIndirectBufferCount => 0;

    // Reusable per-frame buffers.
    private readonly List<int> _visibleSlots = new();
    private readonly HashSet<uint> _visibleCellIds = new();
    private DrawElementsIndirectCommand[] _deicScratch = Array.Empty<DrawElementsIndirectCommand>();

    // Diag.
    public int LoadedSlots   => _alloc.LoadedCount;
    public int VisibleSlots  => _visibleSlots.Count;
    public int CapacitySlots => _alloc.Capacity;
    /// <summary>
    /// Outdoor landcells admitted by the current landscape view. The set is
    /// accumulated across doorway landscape slices and consumed after the
    /// completed render frame by particle visibility.
    /// </summary>
    internal HashSet<uint> VisibleCellIds => _visibleCellIds;

    public void BeginVisibilityFrame() => _visibleCellIds.Clear();

    /// <summary>
    /// Resets the per-GPU-fenced-frame-slot draw state. A retail outside view
    /// may draw terrain more than once in a frame.
    /// </summary>
    public void BeginFrame(int frameSlot)
    {
        ArgumentOutOfRangeException.ThrowIfNegative(frameSlot);
        _retirementLedger.RetryPendingPublications();
        _dynamicFrameSlot = frameSlot;
        _dynamicFrameStarted = true;
    }

    /// <summary>
    /// Two-tier streaming entry point. Accepts a prebuilt mesh from
    /// <see cref="LandblockStreamResult.Loaded.MeshData"/> built on the worker
    /// thread, together with the world-space origin computed by the caller
    /// (render-thread GameWindow derives it from landblockId + liveCenterX/Y).
    ///
    /// Delegates to <see cref="AddLandblock(uint,LandblockMeshData,Vector3)"/>
    /// so both paths share one upload path. Per Phase A.5 spec T15.
    /// </summary>
    public void AddLandblockWithMesh(uint landblockId, LandblockMeshData meshData, Vector3 worldOrigin)
        => AddLandblock(landblockId, meshData, worldOrigin);

    public void AddLandblock(uint landblockId, LandblockMeshData meshData, Vector3 worldOrigin)
    {
        ArgumentNullException.ThrowIfNull(meshData);
        if (meshData.Vertices.Length != VertsPerLandblock)
            throw new ArgumentException(
                $"Expected {VertsPerLandblock} vertices, got {meshData.Vertices.Length}",
                nameof(meshData));
        if (meshData.Indices.Length != IndicesPerLandblock)
            throw new ArgumentException(
                $"Expected {IndicesPerLandblock} indices, got {meshData.Indices.Length}",
                nameof(meshData));

        // A prior replacement may have committed the logical slot switch
        // before queue publication failed. Retry those retained physical-slot
        // transactions before allocating more terrain storage.
        _alloc.RetryPendingPublications();

        bool replacing = _idToSlot.TryGetValue(landblockId, out int replacedSlot);
        int slot = _alloc.Allocate(out var needsGrow);
        bool published = false;
        try
        {
            if (needsGrow)
            {
                int newCap = Math.Max(_alloc.Capacity * 2, slot + 1);
                EnsureCapacity(newCap);
            }

            // Bake worldOrigin into vertex positions; capture min/max Z for AABB.
            var bakedVerts = new TerrainVertex[VertsPerLandblock];
            float zMin = float.MaxValue, zMax = float.MinValue;
            for (int i = 0; i < VertsPerLandblock; i++)
            {
                var v = meshData.Vertices[i];
                var worldPos = v.Position + worldOrigin;
                bakedVerts[i] = new TerrainVertex(worldPos, v.Normal, v.Data0, v.Data1, v.Data2, v.Data3);
                if (worldPos.Z < zMin) zMin = worldPos.Z;
                if (worldPos.Z > zMax) zMax = worldPos.Z;
            }
            if (zMin == float.MaxValue) { zMin = 0f; zMax = 0f; }

            // Bake baseVertex into indices on the CPU side (driver-portable pattern).
            uint baseVertex = (uint)(slot * VertsPerLandblock);
            var bakedIndices = new uint[IndicesPerLandblock];
            for (int i = 0; i < IndicesPerLandblock; i++)
                bakedIndices[i] = meshData.Indices[i] + baseVertex;

            UploadRhiLandblock(slot, bakedVerts, bakedIndices);

            _slots[slot] = new SlotData
            {
                LandblockId = landblockId,
                WorldOrigin = worldOrigin,
                FirstIndex  = (uint)(slot * IndicesPerLandblock),
                IndexCount  = IndicesPerLandblock,
                AabbMin = new Vector3(worldOrigin.X, worldOrigin.Y, zMin),
                AabbMax = new Vector3(worldOrigin.X + LandblockSize, worldOrigin.Y + LandblockSize, zMax),
            };
            _idToSlot[landblockId] = slot;
            published = true;

            if (replacing)
            {
                _slots[replacedSlot] = null;
                _alloc.FreeAfterGpuUse(replacedSlot);
            }
        }
        finally
        {
            if (!published)
                _alloc.ReleaseUnsubmitted(slot);
        }
    }

    public void RemoveLandblock(uint landblockId)
    {
        // Removal clears the logical lookup before retirement publication. A
        // retry therefore has to advance retained publications even when the
        // landblock is no longer present in the map.
        _alloc.RetryPendingPublications();
        if (!_idToSlot.TryGetValue(landblockId, out var slot))
            return;
        _idToSlot.Remove(landblockId);
        _slots[slot] = null;
        _alloc.FreeAfterGpuUse(slot);
        // No GPU clear: the per-frame DEIC array won't reference this slot.
    }

    public void Draw(
        ICamera camera,
        FrustumPlanes? frustum = null,
        uint? neverCullLandblockId = null,
        ReadOnlySpan<Vector4> clipPlanes = default,
        Vector4? ndcClipAabb = null)
    {
        if (_alloc.LoadedCount == 0) return;

        Matrix4x4 viewProjection = camera.View * camera.Projection;

        // Build visible slot list with per-slot frustum cull.
        _visibleSlots.Clear();
        for (int slot = 0; slot < _slots.Length; slot++)
        {
            var data = _slots[slot];
            if (data is null) continue;
            if (frustum is not null && data.LandblockId != neverCullLandblockId)
            {
                if (!FrustumCuller.IsAabbVisible(frustum.Value, data.AabbMin, data.AabbMax))
                    continue;
            }
            _visibleSlots.Add(slot);
            CollectVisibleCells(
                _visibleCellIds,
                data.LandblockId,
                data.WorldOrigin,
                data.AabbMin.Z,
                data.AabbMax.Z,
                frustum,
                viewProjection,
                clipPlanes,
                ndcClipAabb);
        }
        if (_visibleSlots.Count == 0) return;

        BuildIndirectCommands();
        if (!_dynamicFrameStarted)
            throw new InvalidOperationException("BeginFrame must be called before drawing terrain.");
        DrawRhi(viewProjection, _visibleSlots.Count);
    }

    /// <summary>
    /// Builds this frame's <c>DrawElementsIndirectCommand</c> array from the
    /// visible slot list. Pure CPU.
    /// </summary>
    private void BuildIndirectCommands()
    {
        if (_deicScratch.Length < _visibleSlots.Count)
            _deicScratch = new DrawElementsIndirectCommand[Math.Max(_visibleSlots.Count, 64)];
        for (int i = 0; i < _visibleSlots.Count; i++)
        {
            var data = _slots[_visibleSlots[i]]!;
            _deicScratch[i] = new DrawElementsIndirectCommand
            {
                Count         = (uint)data.IndexCount,
                InstanceCount = 1u,
                FirstIndex    = data.FirstIndex,
                BaseVertex    = 0,            // baked into indices on upload
                BaseInstance  = 0,
            };
        }
    }

    public void Dispose()
    {
        if (_disposed)
            return;
        _retirementLedger.RetryPendingPublications();
        DisposeRhi();
    }

    // ----------------------------------------------------------------
    // Private helpers
    // ----------------------------------------------------------------

    internal static void CollectVisibleCells(
        HashSet<uint> destination,
        uint landblockId,
        Vector3 worldOrigin,
        float zMin,
        float zMax,
        FrustumPlanes? frustum,
        Matrix4x4 viewProjection,
        ReadOnlySpan<Vector4> clipPlanes,
        Vector4? ndcClipAabb = null)
    {
        ArgumentNullException.ThrowIfNull(destination);
        const float cellSize = AcDream.Core.Physics.TerrainSurface.CellSize;
        const int cellsPerSide = AcDream.Core.Physics.TerrainSurface.CellsPerSide;
        uint prefix = landblockId & 0xFFFF0000u;

        for (int cellX = 0; cellX < cellsPerSide; cellX++)
        {
            float minX = worldOrigin.X + cellX * cellSize;
            float maxX = minX + cellSize;
            for (int cellY = 0; cellY < cellsPerSide; cellY++)
            {
                float minY = worldOrigin.Y + cellY * cellSize;
                float maxY = minY + cellSize;
                var cellMin = new Vector3(minX, minY, zMin);
                var cellMax = new Vector3(maxX, maxY, zMax);
                if (frustum is not null
                    && !FrustumCuller.IsAabbVisible(frustum.Value, cellMin, cellMax))
                {
                    continue;
                }

                // Retail publishes landcell in_view from the clipped landscape
                // view, not merely from the camera frustum. The modern renderer
                // expresses each doorway slice as homogeneous clip-space planes
                // plus its scissor AABB; use both products here so particle
                // simulation follows the same visible terrain slice as the GPU.
                if (!IsAabbVisibleThroughClipRegion(
                        cellMin,
                        cellMax,
                        viewProjection,
                        clipPlanes,
                        ndcClipAabb))
                {
                    continue;
                }

                uint low = AcDream.Core.Physics.TerrainSurface.ComputeOutdoorCellLowId(
                    cellX * cellSize,
                    cellY * cellSize);
                destination.Add(prefix | low);
            }
        }
    }

    private static bool IsAabbVisibleThroughClipRegion(
        Vector3 min,
        Vector3 max,
        Matrix4x4 viewProjection,
        ReadOnlySpan<Vector4> clipPlanes,
        Vector4? ndcClipAabb)
    {
        Vector4 aabb = ndcClipAabb.GetValueOrDefault();
        bool hasScissorConstraint = ndcClipAabb.HasValue
            && (aabb.X > -1f || aabb.Y > -1f || aabb.Z < 1f || aabb.W < 1f);
        if (clipPlanes.IsEmpty && !hasScissorConstraint)
            return true;

        Span<Vector4> clipCorners = stackalloc Vector4[8];
        for (int corner = 0; corner < clipCorners.Length; corner++)
        {
            var world = new Vector4(
                (corner & 1) == 0 ? min.X : max.X,
                (corner & 2) == 0 ? min.Y : max.Y,
                (corner & 4) == 0 ? min.Z : max.Z,
                1f);
            clipCorners[corner] = Vector4.Transform(world, viewProjection);
        }

        for (int planeIndex = 0; planeIndex < clipPlanes.Length; planeIndex++)
        {
            if (IsAabbOutsideHomogeneousPlane(clipCorners, clipPlanes[planeIndex]))
            {
                return false;
            }
        }

        if (!hasScissorConstraint)
            return true;

        Span<Vector4> scissorPlanes = stackalloc Vector4[4]
        {
            new( 1f,  0f, 0f, -aabb.X),
            new(-1f,  0f, 0f,  aabb.Z),
            new( 0f,  1f, 0f, -aabb.Y),
            new( 0f, -1f, 0f,  aabb.W),
        };
        for (int planeIndex = 0; planeIndex < scissorPlanes.Length; planeIndex++)
        {
            if (IsAabbOutsideHomogeneousPlane(clipCorners, scissorPlanes[planeIndex]))
            {
                return false;
            }
        }

        return true;
    }

    private static bool IsAabbOutsideHomogeneousPlane(
        ReadOnlySpan<Vector4> clipCorners,
        Vector4 plane)
    {
        // A linear half-space reaches its maximum over the transformed AABB at
        // one of the eight corners. If every corner is negative, no point in
        // the cell box can survive this GPU clip plane.
        for (int corner = 0; corner < clipCorners.Length; corner++)
        {
            if (Vector4.Dot(plane, clipCorners[corner]) >= 0f)
                return false;
        }

        return true;
    }

    private void EnsureCapacity(int newCapacity)
    {
        if (newCapacity <= _alloc.Capacity)
            return;
        EnsureRhiCapacity(newCapacity);
    }

    private sealed class SlotData
    {
        public uint    LandblockId;
        public Vector3 WorldOrigin;
        public uint    FirstIndex;
        public int     IndexCount;
        public Vector3 AabbMin;
        public Vector3 AabbMax;
    }
}