ModernUO/Projects/UOContent/Engines/Pathing/Cache/StepCacheFile.cs
Kamron Batman c265bcbb5e
feat(pathfinding): .swb format v7 per-chunk compression (#3a) (#2470)
## Summary
Phase #3a, stacked on #2469. Compresses each chunk record independently with libdeflate (random access preserved).

Trammel: 124.7 MB → 19.2 MB (−85%).

## Details
- Whole-record framing `[u32 UncompressedLen][payload]`; records that don't shrink (tiny Uniform) are stored raw, detected as payload length == UncompressedLen.
- Codec chosen by full-Trammel spike: libdeflate VeryHigh (16.5 MB, 1.83 µs/chunk decompress) over zstd L19/22 (17.4 MB) and managed Brotli q11 (16.4 MB) — best native ratio, fastest decompress, already the repo's packet codec (no new dependency).
- Reuses cached thread-static bindings: `Deflate.Maximum` for bake, `Deflate.Standard` for reads.
- Compression is bake-time only; decompression is one-time per chunk (LRU-cached).
- Format v7; v6 files rejected and re-baked once.

## Tests
v7 unit tests + the v6 suite run through the compression path; full pathfinding suite green; Release build clean.
2026-06-07 00:22:11 -07:00

815 lines
35 KiB
C#
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

using System;
using System.Buffers.Binary;
using System.Collections.Generic;
using System.IO;
using System.IO.Compression;
using System.Runtime.InteropServices;
using Server.Compression;
namespace Server.Engines.Pathing.Cache;
/// <summary>
/// Binary serializer + lazy reader for the step cache. Persists chunk records to disk
/// so a server warm-starts without paying chunk-build cost on the first pathfind through
/// a region. Lazy: opening a file reads only the header + chunk-offset index (~few KB
/// for tens of thousands of chunks), then individual chunks are seeked + deserialized
/// only when the cache asks for them. RAM stays bounded by MaxResidentChunks regardless
/// of file size.
///
/// File layout v7 (little-endian, BufferWriter / BufferReader convention):
///
/// Header (48 bytes):
/// u32 Magic = 0x42575300 ('SWB\0')
/// u32 Version = current FormatVersion (7)
/// u32 MapId
/// u64 Fingerprint XxHash3 over (1) LandTable + ItemTable flags AND (2) the
/// on-disk bytes of mapX.mul / .uop, staidxX.mul, staticsX.mul.
/// Rejects a load when EITHER tile flags shifted (client patch)
/// OR the map data was rewritten (CentredSharp / UOFiddler edit).
/// The .mul format has no built-in CRC; this is the only way
/// to detect those mutations.
/// u64 BakeTimestamp DateTime.UtcNow.Ticks at write time (informational).
/// u32 ChunkCount
/// u64 IndexOffset File position where the chunk index begins.
///
/// Per chunk (ChunkCount times, variable size):
/// u32 UncompressedLen Size of the inflated record body below.
/// byte[] Payload The record body (the v6 layout that follows), libdeflate-
/// compressed. If the on-disk payload length (index recordLength
/// 4) equals UncompressedLen, the body was stored raw because
/// compression did not shrink it (tiny Uniform records).
///
/// Record body (after inflate — the v6 layout):
/// u16 ChunkX
/// u16 ChunkY
/// u32 BuiltMultisVersion
/// u8 Kind 0 = Full; 2 = Uniform
/// // Uniform (Kind == 2): ~28-byte record — all 256 cells share these single values:
/// byte walkMask, wetMask; sbyte sourceZ; sbyte walkZ_N..NW (8); sbyte swimZ_N..NW (8)
/// // Full (Kind == 0) body:
/// u8 HasStrata 0 = single-Z chunk (no strata trailer); 1 = strata trailer follows
/// u8 HasSwimLayer 0 = no shore cells (no swim trailer); 1 = swim trailer follows
/// u16 ZArrayMask bit d set => base directional Z array d is present below as a
/// residual[256] block; cleared => array equals its prediction and is
/// omitted (synthesized at read). bits 0-7 = WalkZ N..NW (predicted via
/// WalkMask), bits 8-15 = SwimZ N..NW (predicted via WetMask).
/// byte WalkMask[256]
/// byte WetMask[256]
/// sbyte SourceZ[256]
/// // For each d in 0..15 with ZArrayMask bit d set, in N,NE,E,SE,S,SW,W,NW order
/// // (walk arrays first, then swim):
/// sbyte residual_d[256] reconstruct: Z_d[c] = (mask bit set ? SourceZ[c] : 0) + residual_d[c]
/// // Swim layer trailer — only when HasSwimLayer == 1 (chunks containing shore cells):
/// sbyte SwimSourceZ[256] (NoSwimLayerCell sentinel = sbyte.MinValue)
/// byte SwimMask[256] (per-cell swim mask baked at SwimSourceZ)
/// sbyte SwimZN_Layer[256]..SwimZNW_Layer[256] (8 arrays, dest-Z at swim perspective)
/// // Strata trailer — only when HasStrata == 1:
/// u16 StrataOffsetByCell[256] (NoStrata sentinel = 0xFFFF)
/// u32 StrataDataLength
/// byte StrataData[StrataDataLength]
/// For each multi-Z cell: u8 stratumCount, then stratumCount × Stratum (19 bytes):
/// sbyte zCenter
/// byte walkMask, wetMask
/// sbyte walkZ_N..NW (8)
/// sbyte swimZ_N..NW (8)
///
/// Index trailer (20 × ChunkCount bytes):
/// For each chunk: { u64 chunkKey, u64 fileOffset, u32 recordLength }
///
/// Per-chunk fixed portion (Kind + flags + ZArrayMask + WalkMask + WetMask + SourceZ):
/// ~783 bytes; each present base Z array adds 256 bytes (0..16 present, so up to ~4 KB).
/// A fully-flat Full chunk stores no residual blocks. Strata trailer: 516 + N × ~30 bytes
/// for a chunk with N multi-Z cells averaging ~2 strata each. LRU bookkeeping
/// (LastTouchedTicks) is intentionally not persisted.
///
/// Files with version &lt; <see cref="MinSupportedVersion"/> are silently rejected
/// at open time (treated as missing) and overwritten on the next save.
/// </summary>
internal static class StepCacheFile
{
public const uint Magic = 0x42575300; // 'SWB\0'
public const uint FormatVersion = 7;
/// <summary>
/// Lowest format version this binary can load. Files below this version are treated as
/// missing (silently rejected) — a subsequent SaveToFile / BakeMap overwrites them with
/// the current FormatVersion. Bumped to 3 when the swim layer landed (v2 had no swim
/// layer). Bumped to 4 when the baker switched to clearance-aware standable-surface
/// strata: v3 bakes anchored every cell at the land average and so missed walkable
/// static-over-land surfaces (sewer/dungeon walkways, bridges, upper building floors),
/// producing ~98% source-Z fallthroughs on those routes. The on-disk layout is
/// unchanged; only the strata population differs, so the bump exists purely to force a
/// one-time re-bake of stale v3 files on first boot under the new binary. Bumped to 5 for
/// uniform-chunk elision: each record now begins with a Kind byte (0 = Full, 2 = Uniform);
/// a fully-uniform chunk (no strata, no swim layer, all 19 base arrays constant) stores
/// one cell's worth of data (~28 bytes) instead of the full record. Bumped to 6 for
/// predictive-Z residuals: each base directional Z array is stored as a masked residual
/// against SourceZ (a ZArrayMask u16 flags which arrays are present); arrays matching their
/// prediction are omitted and synthesized at read. v5 files are rejected and re-baked once.
/// Bumped to 7 for per-chunk compression: each chunk record is libdeflate-compressed
/// independently (random access preserved) behind a u32 uncompressed-length prefix; tiny
/// records that do not shrink are stored raw. v6 files are rejected and re-baked once.
/// </summary>
public const uint MinSupportedVersion = 7;
// Per-chunk record discriminator (first byte after BuiltMultisVersion). 1 is reserved.
private const byte KindFull = 0;
private const byte KindUniform = 2;
private const int HeaderSize =
sizeof(uint) // Magic
+ sizeof(uint) // Version
+ sizeof(uint) // MapId
+ sizeof(ulong) // Fingerprint
+ sizeof(ulong) // BakeTimestamp
+ sizeof(uint) // ChunkCount
+ sizeof(ulong); // IndexOffset
// Index entry: chunkKey + fileOffset + recordLength. Bumped to include length when
// strata made chunk records variable-size; the lazy reader uses length to do a
// single bulk read per chunk without consulting the next offset.
private const int IndexEntryBytes = sizeof(ulong) + sizeof(ulong) + sizeof(uint);
/// <summary>Fixed-size portion of a chunk record (everything except the optional strata + swim trailers).</summary>
private const int BytesPerChunkBase =
sizeof(ushort) + sizeof(ushort) + sizeof(uint)
+ sizeof(byte) + sizeof(byte) + sizeof(byte) // Kind + HasStrata + HasSwimLayer
+ sizeof(ushort) // ZArrayMask
+ StepChunk.CellsPerChunk // WalkMask
+ StepChunk.CellsPerChunk // WetMask
+ StepChunk.CellsPerChunk // SourceZ
+ 8 * StepChunk.CellsPerChunk // WalkZ[8]
+ 8 * StepChunk.CellsPerChunk; // SwimZ[8]
/// <summary>Swim-layer trailer overhead when present: per-cell SourceZ + Mask + 8×Z arrays.</summary>
private const int SwimLayerOverhead = 10 * StepChunk.CellsPerChunk;
/// <summary>Strata trailer overhead when present: 256×u16 offset table + u32 data length.</summary>
private const int StrataTrailerOverhead = StepChunk.CellsPerChunk * sizeof(ushort) + sizeof(uint);
/// <summary>
/// Byte offset of the IndexOffset u64 within the header
/// (Magic+Version+MapId+Fingerprint+BakeTimestamp+ChunkCount = 32). Patched after chunks land.
/// </summary>
private const int IndexOffsetFieldPosition = 32;
public delegate bool ChunkEnumerator(out int chunkX, out int chunkY, out StepChunk chunk);
/// <summary>
/// Peek at a .swb file's Fingerprint field (header byte offset 12) without
/// reading any chunk data. Returns false on missing file, bad magic, or wrong
/// version. Cheap — reads 20 bytes total.
/// </summary>
public static bool TryReadFingerprint(string path, out ulong fingerprint)
{
fingerprint = 0;
if (!File.Exists(path))
{
return false;
}
try
{
using var stream = new FileStream(path, FileMode.Open, FileAccess.Read, FileShare.Read | FileShare.Delete);
Span<byte> buf = stackalloc byte[20];
if (stream.Read(buf) < 20)
{
return false;
}
if (BinaryPrimitives.ReadUInt32LittleEndian(buf) != Magic)
{
return false;
}
var version = BinaryPrimitives.ReadUInt32LittleEndian(buf[4..]);
if (version < MinSupportedVersion || version > FormatVersion)
{
return false;
}
// mapId is at buf[8..12], we skip; hash is at buf[12..20].
fingerprint = BinaryPrimitives.ReadUInt64LittleEndian(buf[12..]);
return true;
}
catch
{
return false;
}
}
/// <summary>
/// Combined XxHash3 fingerprint over (1) the loaded TileData flag tables and (2) the
/// per-map .mul / .uop file contents (via <see cref="TileMatrix.MapFilesFingerprint"/>).
/// Bake files carry this hash so a load can refuse to populate the cache when EITHER
/// tile flags shifted (client patch) OR the map data was rewritten (CentredSharp /
/// UOFiddler edit). The .mul format has no built-in CRC; this is the only way to
/// detect those mutations.
/// </summary>
public static ulong ComputeFingerprint(int mapId)
{
var hasher = HashUtility.CreateXxHash3();
// TileData flag tables — same projection trick as before: just the Flags ulong
// from each entry, written little-endian into a contiguous byte buffer. The
// struct itself has a string Name (reference) whose object identity isn't
// stable across runs, so MemoryMarshal.Cast over the whole struct would drift.
var landTable = TileData.LandTable;
var itemTable = TileData.ItemTable;
var bytes = new byte[(landTable.Length + itemTable.Length) * sizeof(ulong)];
var span = bytes.AsSpan();
for (var i = 0; i < landTable.Length; i++)
{
BinaryPrimitives.WriteUInt64LittleEndian(span[(i * 8)..], (ulong)landTable[i].Flags);
}
var itemOffset = landTable.Length * 8;
for (var i = 0; i < itemTable.Length; i++)
{
BinaryPrimitives.WriteUInt64LittleEndian(span[(itemOffset + i * 8)..], (ulong)itemTable[i].Flags);
}
hasher.Append(bytes);
// Map files (mapX.mul / .uop, staidxX.mul, staticsX.mul). TileMatrix already
// streamed them through XxHash3 once at construction; mix the result in.
var map = Map.Maps[mapId];
if (map != null && map != Map.Internal && map.Tiles != null)
{
Span<byte> mapHashBytes = stackalloc byte[sizeof(ulong)];
BinaryPrimitives.WriteUInt64LittleEndian(mapHashBytes, map.Tiles.MapFilesFingerprint);
hasher.Append(mapHashBytes);
}
return hasher.GetCurrentHashAsUInt64();
}
/// <summary>
/// Writes the file: header (with placeholder IndexOffset) → chunks (offsets recorded)
/// → index trailer → patches the header IndexOffset. <paramref name="chunkCount"/> must
/// equal the actual number of chunks <paramref name="next"/> will yield.
/// </summary>
public static void Write(string path, uint mapId, uint chunkCount, ChunkEnumerator next)
{
Directory.CreateDirectory(Path.GetDirectoryName(path) ?? ".");
// Initial estimate: base record + a modest strata budget per chunk. Coastline
// chunks add another ~2.5 KB (swim layer) but they're a small fraction of any
// map; the writer grows on overflow so under-estimating just causes a few
// realloc/copy cycles during the bake — not a correctness issue.
var capacity = HeaderSize
+ (BytesPerChunkBase + 256) * (int)chunkCount
+ IndexEntryBytes * (int)chunkCount;
var buffer = new byte[capacity];
var w = new BufferWriter(buffer, prefixStr: false);
w.Write(Magic);
w.Write(FormatVersion);
w.Write(mapId);
w.Write(ComputeFingerprint((int)mapId));
w.Write((ulong)DateTime.UtcNow.Ticks);
w.Write(chunkCount);
w.Write(0UL); // IndexOffset placeholder, patched after chunks
// Per-chunk compression: each record is built uncompressed into `recordScratch`, then
// libdeflate-compressed into `compScratch` and framed as [u32 uncompressedLen][payload].
// Reuse the cached per-thread VeryHigh compressor (construction allocates native state).
// libdeflate is not thread-safe, but bake runs single-threaded. VeryHigh is the best-ratio
// level and its slow compression is irrelevant offline (decompression is what's hot, ~1.8us).
var packer = Deflate.Maximum;
var recordScratch = new byte[BytesPerChunkBase + 1024];
var compScratch = new byte[packer.MaxPackSize(recordScratch.Length)];
var indexEntries = new (ulong key, ulong offset, uint length)[chunkCount];
var written = 0u;
while (next(out var chunkX, out var chunkY, out var chunk))
{
if (written >= chunkCount)
{
throw new InvalidOperationException(
$"StepCacheFile.Write: enumerator yielded more than the declared {chunkCount} chunks"
);
}
var chunkOffset = (ulong)w.Position;
WriteChunk(w, chunkX, chunkY, chunk, packer, ref recordScratch, ref compScratch);
var chunkLength = (uint)((ulong)w.Position - chunkOffset);
indexEntries[written] = (PackChunkKey(chunkX, chunkY), chunkOffset, chunkLength);
written++;
}
if (written != chunkCount)
{
throw new InvalidOperationException(
$"StepCacheFile.Write: declared {chunkCount} chunks but enumerator yielded {written}"
);
}
var indexOffset = (ulong)w.Position;
for (var i = 0u; i < chunkCount; i++)
{
w.Write(indexEntries[i].key);
w.Write(indexEntries[i].offset);
w.Write(indexEntries[i].length);
}
// Patch IndexOffset on the writer's current backing buffer (BufferWriter may
// have grown during chunk writes; the original `buffer` ref is stale after grow).
var liveBuffer = w.Buffer;
BinaryPrimitives.WriteUInt64LittleEndian(liveBuffer.AsSpan(IndexOffsetFieldPosition, 8), indexOffset);
var totalBytes = (int)w.Position;
File.WriteAllBytes(path, liveBuffer.AsSpan(0, totalBytes).ToArray());
}
/// <summary>
/// Opens a .swb file and reads only its header + chunk-offset index. Returns null on
/// missing file, magic / version mismatch, or Fingerprint mismatch (a stale bake
/// against a freshly patched client). Callers own disposal of the returned reader.
/// </summary>
public static LazyReader OpenForLazy(string path)
{
if (!File.Exists(path))
{
return null;
}
FileStream stream = null;
try
{
stream = new FileStream(
path,
FileMode.Open,
FileAccess.Read,
FileShare.Read | FileShare.Delete
);
Span<byte> headerBuf = stackalloc byte[HeaderSize];
if (stream.Read(headerBuf) != HeaderSize)
{
stream.Dispose();
return null;
}
var magic = BinaryPrimitives.ReadUInt32LittleEndian(headerBuf);
if (magic != Magic)
{
stream.Dispose();
return null;
}
var version = BinaryPrimitives.ReadUInt32LittleEndian(headerBuf[4..]);
if (version < MinSupportedVersion || version > FormatVersion)
{
// Below the minimum supported version: treat as missing. Older files
// get silently overwritten on the next SaveToFile / BakeMap.
stream.Dispose();
return null;
}
var mapId = BinaryPrimitives.ReadUInt32LittleEndian(headerBuf[8..]);
var fingerprint = BinaryPrimitives.ReadUInt64LittleEndian(headerBuf[12..]);
var bakeTimestamp = BinaryPrimitives.ReadUInt64LittleEndian(headerBuf[20..]);
var chunkCount = BinaryPrimitives.ReadUInt32LittleEndian(headerBuf[28..]);
var indexOffset = BinaryPrimitives.ReadUInt64LittleEndian(headerBuf[32..]);
if (fingerprint != ComputeFingerprint((int)mapId))
{
stream.Dispose();
return null;
}
// Read the chunk-offset index in one shot.
var indexBytes = (int)chunkCount * IndexEntryBytes;
var indexBuf = new byte[indexBytes];
stream.Position = (long)indexOffset;
if (stream.Read(indexBuf, 0, indexBytes) != indexBytes)
{
stream.Dispose();
return null;
}
var offsets = new Dictionary<ulong, (ulong offset, uint length)>((int)chunkCount);
for (var i = 0; i < chunkCount; i++)
{
var entry = indexBuf.AsSpan(i * IndexEntryBytes);
var key = BinaryPrimitives.ReadUInt64LittleEndian(entry);
var off = BinaryPrimitives.ReadUInt64LittleEndian(entry[8..]);
var len = BinaryPrimitives.ReadUInt32LittleEndian(entry[16..]);
offsets[key] = (off, len);
}
return new LazyReader(stream, mapId, fingerprint, bakeTimestamp, chunkCount, offsets);
}
catch
{
stream?.Dispose();
return null;
}
}
private static ulong PackChunkKey(int chunkX, int chunkY) => ((ulong)(uint)chunkX << 32) | (uint)chunkY;
/// <summary>
/// Predicted directional-Z for one cell/direction: the cell's own SourceZ when the
/// direction is walkable/wet (mask bit set), else 0 — matching the baker, which leaves
/// non-walkable directional slots at their zero-initialized default
/// (StepProbe.ComputeMaskAt clears walkZs/swimZs and writes only on a successful step).
/// </summary>
internal static sbyte Predict(byte dirMaskByte, int bit, sbyte sourceZ) =>
(dirMaskByte >> bit & 1) != 0 ? sourceZ : (sbyte)0;
/// <summary>
/// Residual of an absolute directional-Z against its prediction. Unchecked two's-complement
/// so the transform is byte-exact for ALL sbyte inputs (no value-range constraint).
/// </summary>
internal static sbyte EncodeResidual(sbyte z, sbyte predict) => unchecked((sbyte)(z - predict));
/// <summary>Inverse of <see cref="EncodeResidual"/>: absolute directional-Z = predict + residual.</summary>
internal static sbyte DecodeZ(sbyte predict, sbyte residual) => unchecked((sbyte)(predict + residual));
/// <summary>
/// The 16 base directional-Z arrays in canonical order: walk N..NW (indices 0-7),
/// then swim N..NW (8-15). Index d uses WalkMask (d &lt; 8) or WetMask (d &gt;= 8)
/// with direction bit (d &amp; 7). Allocates a 16-slot reference array (bake/read time only).
/// </summary>
private static sbyte[][] GetBaseZArrays(StepChunk c) => new[]
{
c.WalkZN, c.WalkZNE, c.WalkZE, c.WalkZSE, c.WalkZS, c.WalkZSW, c.WalkZW, c.WalkZNW,
c.SwimZN, c.SwimZNE, c.SwimZE, c.SwimZSE, c.SwimZS, c.SwimZSW, c.SwimZW, c.SwimZNW,
};
/// <summary>
/// Builds the uncompressed v6 record for one chunk into <paramref name="w"/>, libdeflate-
/// compresses it, and writes it framed as [u32 uncompressedLen][payload]. The payload is the
/// compressed bytes, or — when compression does not shrink the record (tiny Uniform records) —
/// the raw record itself; the reader distinguishes the two by payload length vs uncompressedLen.
/// </summary>
private static void WriteChunk(
BufferWriter w, int chunkX, int chunkY, StepChunk chunk,
LibDeflateBinding packer, ref byte[] recordScratch, ref byte[] compScratch
)
{
var rw = new BufferWriter(recordScratch, prefixStr: false);
BuildRecord(rw, chunkX, chunkY, chunk);
recordScratch = rw.Buffer; // may have grown; keep the larger buffer for reuse
var recordLen = (int)rw.Position;
var bound = packer.MaxPackSize(recordLen);
if (compScratch.Length < bound)
{
compScratch = new byte[bound];
}
var compLen = packer.Pack(compScratch, recordScratch.AsSpan(0, recordLen));
w.Write((uint)recordLen);
if (compLen > 0 && compLen < recordLen)
{
w.Write(compScratch.AsSpan(0, compLen));
}
else
{
// Incompressible (or expanded): store the record raw. The reader detects this when
// the on-disk payload length equals the uncompressed length.
w.Write(recordScratch.AsSpan(0, recordLen));
}
}
private static void BuildRecord(BufferWriter w, int chunkX, int chunkY, StepChunk chunk)
{
w.Write((ushort)chunkX);
w.Write((ushort)chunkY);
w.Write((uint)chunk.BuiltMultisVersion);
// Kind: 0 = Full, 2 = Uniform. A uniform chunk (no strata, no swim layer, all 19 base
// arrays constant) stores one cell's worth of data (~28-byte record total).
if (chunk.IsUniform())
{
w.Write(KindUniform);
w.Write(chunk.WalkMask[0]);
w.Write(chunk.WetMask[0]);
w.Write((byte)chunk.SourceZ[0]);
w.Write((byte)chunk.WalkZN[0]);
w.Write((byte)chunk.WalkZNE[0]);
w.Write((byte)chunk.WalkZE[0]);
w.Write((byte)chunk.WalkZSE[0]);
w.Write((byte)chunk.WalkZS[0]);
w.Write((byte)chunk.WalkZSW[0]);
w.Write((byte)chunk.WalkZW[0]);
w.Write((byte)chunk.WalkZNW[0]);
w.Write((byte)chunk.SwimZN[0]);
w.Write((byte)chunk.SwimZNE[0]);
w.Write((byte)chunk.SwimZE[0]);
w.Write((byte)chunk.SwimZSE[0]);
w.Write((byte)chunk.SwimZS[0]);
w.Write((byte)chunk.SwimZSW[0]);
w.Write((byte)chunk.SwimZW[0]);
w.Write((byte)chunk.SwimZNW[0]);
return;
}
w.Write(KindFull); // Full
var strataOffsetByCell = chunk.GetStrataOffsetByCellForSerialization();
var strataData = chunk.GetStrataDataForSerialization();
var hasStrata = strataOffsetByCell != null;
var hasSwimLayer = chunk.HasSwimLayer;
w.Write((byte)(hasStrata ? 1 : 0));
w.Write((byte)(hasSwimLayer ? 1 : 0));
// Predictive-Z: each base directional Z array is stored as a masked residual against
// SourceZ. Bit d of ZArrayMask is set only when array d differs from its prediction
// somewhere; cleared arrays are omitted and rebuilt from mask+SourceZ at read.
var zArrays = GetBaseZArrays(chunk);
ushort zArrayMask = 0;
for (var d = 0; d < 16; d++)
{
var z = zArrays[d];
var dirMask = d < 8 ? chunk.WalkMask : chunk.WetMask;
var bit = d & 7;
for (var cell = 0; cell < StepChunk.CellsPerChunk; cell++)
{
if (z[cell] != Predict(dirMask[cell], bit, chunk.SourceZ[cell]))
{
zArrayMask |= (ushort)(1 << d);
break;
}
}
}
w.Write(zArrayMask);
w.Write(chunk.WalkMask);
w.Write(chunk.WetMask);
WriteSBytes(w, chunk.SourceZ);
Span<sbyte> residual = stackalloc sbyte[StepChunk.CellsPerChunk];
for (var d = 0; d < 16; d++)
{
if ((zArrayMask >> d & 1) == 0)
{
continue;
}
var z = zArrays[d];
var dirMask = d < 8 ? chunk.WalkMask : chunk.WetMask;
var bit = d & 7;
for (var cell = 0; cell < StepChunk.CellsPerChunk; cell++)
{
residual[cell] = EncodeResidual(z[cell], Predict(dirMask[cell], bit, chunk.SourceZ[cell]));
}
w.Write(MemoryMarshal.Cast<sbyte, byte>(residual));
}
if (hasSwimLayer)
{
WriteSBytes(w, chunk.SwimSourceZ);
w.Write(chunk.SwimMask);
WriteSBytes(w, chunk.SwimZN_Layer);
WriteSBytes(w, chunk.SwimZNE_Layer);
WriteSBytes(w, chunk.SwimZE_Layer);
WriteSBytes(w, chunk.SwimZSE_Layer);
WriteSBytes(w, chunk.SwimZS_Layer);
WriteSBytes(w, chunk.SwimZSW_Layer);
WriteSBytes(w, chunk.SwimZW_Layer);
WriteSBytes(w, chunk.SwimZNW_Layer);
}
if (hasStrata)
{
// 256 × u16 offsets, then u32 length-prefixed strata byte array.
for (var i = 0; i < StepChunk.CellsPerChunk; i++)
{
w.Write(strataOffsetByCell[i]);
}
var dataLen = (uint)(strataData?.Length ?? 0);
w.Write(dataLen);
if (dataLen > 0)
{
w.Write(strataData);
}
}
}
private static StepChunk ReadChunk(byte[] buffer)
{
var r = new BufferReader(buffer);
// Skip ChunkX + ChunkY (already known via the index lookup).
r.ReadUShort();
r.ReadUShort();
var multisVersion = (int)r.ReadUInt();
var kind = r.ReadByte();
var chunk = new StepChunk { BuiltMultisVersion = multisVersion };
if (kind == KindUniform) // Uniform — one cell's worth of the 19 base arrays, fill all 256 cells.
{
Array.Fill(chunk.WalkMask, r.ReadByte());
Array.Fill(chunk.WetMask, r.ReadByte());
Array.Fill(chunk.SourceZ, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZN, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZNE, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZE, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZSE, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZS, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZSW, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZW, (sbyte)r.ReadByte());
Array.Fill(chunk.WalkZNW, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZN, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZNE, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZE, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZSE, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZS, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZSW, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZW, (sbyte)r.ReadByte());
Array.Fill(chunk.SwimZNW, (sbyte)r.ReadByte());
return chunk;
}
var hasStrata = r.ReadByte() != 0;
var hasSwimLayer = r.ReadByte() != 0;
var zArrayMask = r.ReadUShort();
r.Read(chunk.WalkMask);
r.Read(chunk.WetMask);
ReadSBytes(r, chunk.SourceZ);
// Predictive-Z reconstruction: present arrays carry residuals (z = predict + residual);
// absent arrays are synthesized from mask+SourceZ (z = predict, residual implicitly 0).
var zArrays = GetBaseZArrays(chunk);
Span<sbyte> residual = stackalloc sbyte[StepChunk.CellsPerChunk];
for (var d = 0; d < 16; d++)
{
var z = zArrays[d];
var dirMask = d < 8 ? chunk.WalkMask : chunk.WetMask;
var bit = d & 7;
if ((zArrayMask >> d & 1) != 0)
{
r.Read(MemoryMarshal.Cast<sbyte, byte>(residual));
for (var cell = 0; cell < StepChunk.CellsPerChunk; cell++)
{
z[cell] = DecodeZ(Predict(dirMask[cell], bit, chunk.SourceZ[cell]), residual[cell]);
}
}
else
{
for (var cell = 0; cell < StepChunk.CellsPerChunk; cell++)
{
z[cell] = Predict(dirMask[cell], bit, chunk.SourceZ[cell]);
}
}
}
if (hasSwimLayer)
{
chunk.AllocateSwimLayer();
ReadSBytes(r, chunk.SwimSourceZ);
r.Read(chunk.SwimMask);
ReadSBytes(r, chunk.SwimZN_Layer);
ReadSBytes(r, chunk.SwimZNE_Layer);
ReadSBytes(r, chunk.SwimZE_Layer);
ReadSBytes(r, chunk.SwimZSE_Layer);
ReadSBytes(r, chunk.SwimZS_Layer);
ReadSBytes(r, chunk.SwimZSW_Layer);
ReadSBytes(r, chunk.SwimZW_Layer);
ReadSBytes(r, chunk.SwimZNW_Layer);
}
if (hasStrata)
{
var offsets = new ushort[StepChunk.CellsPerChunk];
for (var i = 0; i < offsets.Length; i++)
{
offsets[i] = r.ReadUShort();
}
var dataLen = (int)r.ReadUInt();
var data = new byte[dataLen];
if (dataLen > 0)
{
r.Read(data);
}
chunk.SetStrata(offsets, data);
}
return chunk;
}
private static void WriteSBytes(BufferWriter w, sbyte[] arr) =>
w.Write(MemoryMarshal.Cast<sbyte, byte>(arr.AsSpan()));
private static void ReadSBytes(BufferReader r, sbyte[] arr) =>
r.Read(MemoryMarshal.Cast<sbyte, byte>(arr.AsSpan()));
/// <summary>
/// Open handle on a .swb file. Holds the FileStream + chunk-offset index. Chunks are
/// fetched on demand via <see cref="TryReadChunk"/>; only the records actually queried
/// are ever materialized. Dispose releases the underlying stream.
/// </summary>
internal sealed class LazyReader : IDisposable
{
private FileStream _stream;
private readonly Dictionary<ulong, (ulong offset, uint length)> _offsets;
private byte[] _buffer; // raw on-disk record: [u32 uncompressedLen][payload]
private byte[] _bodyBuffer; // decompressed v6 record, parsed by ReadChunk
public uint MapId { get; }
public ulong Fingerprint { get; }
public ulong BakeTimestamp { get; }
public uint ChunkCount { get; }
public int IndexedChunkCount => _offsets.Count;
public bool Has(int chunkX, int chunkY) => _offsets.ContainsKey(PackChunkKey(chunkX, chunkY));
/// <summary>
/// Enumerates every (chunkX, chunkY) coordinate the file holds. Used by
/// <see cref="StepCache"/> when preload is enabled to materialize all chunks
/// upfront instead of on first query.
/// </summary>
public IEnumerable<(int chunkX, int chunkY)> EnumerateChunkCoords()
{
foreach (var key in _offsets.Keys)
{
yield return ((int)(key >> 32), (int)(key & 0xFFFFFFFF));
}
}
internal LazyReader(
FileStream stream, uint mapId, ulong fingerprint, ulong bakeTimestamp,
uint chunkCount, Dictionary<ulong, (ulong offset, uint length)> offsets
)
{
_stream = stream;
MapId = mapId;
Fingerprint = fingerprint;
BakeTimestamp = bakeTimestamp;
ChunkCount = chunkCount;
_offsets = offsets;
_buffer = new byte[BytesPerChunkBase];
_bodyBuffer = new byte[BytesPerChunkBase];
}
/// <summary>
/// Returns the chunk record at (<paramref name="chunkX"/>, <paramref name="chunkY"/>)
/// from the file, or null if the file doesn't contain it. Single seek + bulk read,
/// sized exactly to the chunk's recorded length (which varies with strata size).
/// </summary>
public StepChunk TryReadChunk(int chunkX, int chunkY)
{
if (_stream == null)
{
return null;
}
var key = PackChunkKey(chunkX, chunkY);
if (!_offsets.TryGetValue(key, out var entry))
{
return null;
}
// Grow the on-disk scratch buffer if this chunk's record is larger than what we have.
if (entry.length > _buffer.Length)
{
_buffer = new byte[entry.length];
}
_stream.Position = (long)entry.offset;
var read = _stream.Read(_buffer, 0, (int)entry.length);
if (read < (int)entry.length || entry.length < sizeof(uint))
{
return null;
}
// Frame: [u32 uncompressedLen][payload]. payload is libdeflate-compressed, unless its
// length equals uncompressedLen, in which case it was stored raw (incompressible).
var uncompressedLen = (int)BinaryPrimitives.ReadUInt32LittleEndian(_buffer);
var payloadLen = (int)entry.length - sizeof(uint);
if (_bodyBuffer.Length < uncompressedLen)
{
_bodyBuffer = new byte[uncompressedLen];
}
if (payloadLen == uncompressedLen)
{
Array.Copy(_buffer, sizeof(uint), _bodyBuffer, 0, uncompressedLen);
}
else
{
// Decompression is level-independent, so reuse the shared per-thread binding
// rather than allocating a native decompressor per reader. The cache is read on
// the single game thread; libdeflate's non-thread-safety is satisfied by ThreadStatic.
var result = Deflate.Standard.Unpack(
_bodyBuffer.AsSpan(0, uncompressedLen),
_buffer.AsSpan(sizeof(uint), payloadLen),
out var produced
);
if (result != LibDeflateResult.Success || produced != uncompressedLen)
{
return null;
}
}
return ReadChunk(_bodyBuffer);
}
public void Dispose()
{
_stream?.Dispose();
_stream = null;
_buffer = null;
_bodyBuffer = null;
}
}
}