update region file format 2 to 3 (WIP)

doc/specs/outdated/region_file_spec_v2.md

@@ -0,0 +1,41 @@
+# Region File (version 2)
+
+File format BNF (RFC 5234):
+
+```bnf
+file    = header (*chunk) offsets   complete file
+header  = magic %x02 %x00           magic number, version and reserved
+                                    zero byte
+
+magic   = %x2E %x56 %x4F %x58       '.VOXREG\0'
+          %x52 %x45 %x47 %x00
+
+chunk   = int32 (*byte)             byte array with size prefix
+offsets = (1024*int32)              offsets table
+int32   = 4byte                     signed big-endian 32 bit integer
+byte    = %x00-FF                   8 bit unsigned integer
+```
+
+C struct visualization:
+
+```c
+typedef unsigned char byte;
+
+struct file {
+	// 10 bytes
+	struct {
+		char magic[8] = ".VOXREG";
+		byte version = 2;
+		byte reserved = 0;
+	} header;
+	
+	struct {
+		int32_t size; // byteorder: big-endian
+		byte* data;
+	} chunks[1024]; // file does not contain zero sizes for missing chunks
+	
+	int32_t offsets[1024]; // byteorder: big-endian
+};
+```
+
+Offsets table contains chunks positions in file. 0 means that chunk is not present in the file. Minimal valid offset is 10 (header size).

doc/specs/outdated/region_voxels_chunk_spec_v1.md

@@ -0,0 +1,26 @@
+# Voxels Chunk (version 1)
+
+Voxel regions layer chunk structure.
+
+Values are separated for extRLE8 compression efficiency.
+
+File format BNF (RFC 5234):
+
+```bnf
+chunk    = (65536*byte)    block indices (most significant bytes)
+           (65536*byte)    block indices (least significant bytes)
+           (65536*byte)    block states (most significant bytes)
+           (65536*byte)    block states (least significant bytes)
+
+byte     = %x00-FF         8 bit unsigned integer
+```
+
+65536 is number of voxels per chunk (16\*256\*16)
+
+## Block state
+
+Block state is encoded in 16 bits:
+- 0-2 bits (3) - block rotation index
+- 3-5 bits (3) - segment block bits
+- 6-7 bits (2) - reserved
+- 8-15 bits (8) - user bits

doc/specs/region_file_spec.md

@@ -1,18 +1,21 @@
-# Region File (version 2)
+# Region File (version 3)
 
 File format BNF (RFC 5234):
 
 ```bnf
 file    = header (*chunk) offsets   complete file
-header  = magic %x02 %x00           magic number, version and reserved
-                                    zero byte
+header  = magic %x02 byte           magic number, version and compression
+                                    method
 
 magic   = %x2E %x56 %x4F %x58       '.VOXREG\0'
           %x52 %x45 %x47 %x00
 
-chunk   = int32 (*byte)             byte array with size prefix
-offsets = (1024*int32)              offsets table
-int32   = 4byte                     signed big-endian 32 bit integer
+chunk   = uint32 uint32 (*byte)     byte array with size and source size 
+                                    prefix where source size is 
+                                    decompressed chunk data size
+
+offsets = (1024*uint32)             offsets table
+int32   = 4byte                     unsigned big-endian 32 bit integer
 byte    = %x00-FF                   8 bit unsigned integer
 ```
 
@@ -25,17 +28,23 @@ struct file {
 	// 10 bytes
 	struct {
 		char magic[8] = ".VOXREG";
-		byte version = 2;
-		byte reserved = 0;
+		byte version = 3;
+		byte compression;
 	} header;
 	
 	struct {
-		int32_t size; // byteorder: big-endian
+		uint32_t size; // byteorder: little-endian
+		uint32_t sourceSize; // byteorder: little-endian
 		byte* data;
 	} chunks[1024]; // file does not contain zero sizes for missing chunks
 	
-	int32_t offsets[1024]; // byteorder: big-endian
+	uint32_t offsets[1024]; // byteorder: little-endian
 };
 ```
 
 Offsets table contains chunks positions in file. 0 means that chunk is not present in the file. Minimal valid offset is 10 (header size).
+
+Available compression methods:
+0. no compression
+1. extRLE8
+2. extRLE16

doc/specs/region_voxels_chunk_spec.md

@@ -1,17 +1,14 @@
-# Voxels Chunk (version 1)
+# Voxels Chunk (version 2)
 
-Voxel regions layer chunk structure.
-
-Values are separated for extRLE8 compression efficiency.
+IDs and states are separated for extRLE16 compression efficiency.
 
 File format BNF (RFC 5234):
 
 ```bnf
-chunk    = (65536*byte)    block indices (most significant bytes)
-           (65536*byte)    block indices (least significant bytes)
-           (65536*byte)    block states (most significant bytes)
-           (65536*byte)    block states (least significant bytes)
+chunk    = (65536*uint16)  block ids
+           (65536*uint16)  block states
 
+uint16   = 2byte           16 bit little-endian unsigned integer
 byte     = %x00-FF         8 bit unsigned integer
 ```
 

src/coders/compression.cpp

@@ -25,29 +25,38 @@ static std::shared_ptr<ubyte[]> get_buffer(size_t minSize) {
     return nullptr;
 }
 
+static auto compress_rle(
+    const ubyte* src,
+    size_t srclen,
+    size_t& len,
+    size_t(*encodefunc)(const ubyte*, size_t, ubyte*)
+) {
+    auto buffer = get_buffer(srclen * 2);
+    auto bytes = buffer.get();
+    std::unique_ptr<ubyte[]> uptr;
+    if (bytes == nullptr) {
+        uptr = std::make_unique<ubyte[]>(srclen * 2);
+        bytes = uptr.get();
+    }
+    len = encodefunc(src, srclen, bytes);
+    if (uptr) {
+        return uptr;
+    }
+    auto data = std::make_unique<ubyte[]>(len);
+    std::memcpy(data.get(), bytes, len);
+    return data;
+}
+
 std::unique_ptr<ubyte[]> compression::compress(
     const ubyte* src, size_t srclen, size_t& len, Method method
 ) {
     switch (method) {
         case Method::NONE:
             throw std::invalid_argument("compression method is NONE");
-        case Method::EXTRLE8: {
-            // max extrle out size is srcLen * 2
-            auto buffer = get_buffer(srclen * 2);
-            auto bytes = buffer.get();
-            std::unique_ptr<ubyte[]> uptr;
-            if (bytes == nullptr) {
-                uptr = std::make_unique<ubyte[]>(srclen * 2);
-                bytes = uptr.get();
-            }
-            len = extrle::encode(src, srclen, bytes);
-            if (uptr) {
-                return uptr;
-            }
-            auto data = std::make_unique<ubyte[]>(len);
-            std::memcpy(data.get(), bytes, len);
-            return data;
-        }
+        case Method::EXTRLE8:
+            return compress_rle(src, srclen, len, extrle::encode);
+        case Method::EXTRLE16:
+            return compress_rle(src, srclen, len, extrle::encode16);
         case Method::GZIP: {
             auto buffer = gzip::compress(src, srclen);
             auto data = std::make_unique<ubyte[]>(buffer.size());
@@ -71,6 +80,11 @@ std::unique_ptr<ubyte[]> compression::decompress(
             extrle::decode(src, srclen, decompressed.get());
             return decompressed;
         }
+        case Method::EXTRLE16: {
+            auto decompressed = std::make_unique<ubyte[]>(dstlen);
+            extrle::decode16(src, srclen, decompressed.get());
+            return decompressed;
+        }
         case Method::GZIP: {
             auto buffer = gzip::decompress(src, srclen);
             if (buffer.size() != dstlen) {

src/coders/compression.hpp

@@ -6,7 +6,7 @@
 
 namespace compression {
     enum class Method {
-        NONE, EXTRLE8, GZIP
+        NONE, EXTRLE8, EXTRLE16, GZIP
     };
 
     /// @brief Compress buffer

src/files/RegionsLayer.cpp

@@ -135,6 +135,10 @@ WorldRegion* RegionsLayer::getRegion(int x, int z) {
     return found->second.get();
 }
 
+fs::path RegionsLayer::getRegionFilePath(int x, int z) const {
+    return folder / get_region_filename(x, z);
+}
+
 WorldRegion* RegionsLayer::getOrCreateRegion(int x, int z) {
     if (auto region = getRegion(x, z)) {
         return region;

src/files/WorldConverter.cpp

@@ -6,6 +6,7 @@
 #include <utility>
 
 #include "content/ContentReport.hpp"
+#include "files/compatibility.hpp"
 #include "data/dynamic.hpp"
 #include "debug/Logger.hpp"
 #include "files/files.hpp"
@@ -48,7 +49,7 @@ void WorldConverter::addRegionsTasks(
             logger.error() << "could not parse region name " << name;
             continue;
         }
-        tasks.push(ConvertTask {taskType, file.path(), x, z});
+        tasks.push(ConvertTask {taskType, file.path(), x, z, layerid});
     }
 }
 
@@ -58,11 +59,7 @@ void WorldConverter::createUpgradeTasks() {
         if (issue.issueType != ContentIssueType::REGION_FORMAT_UPDATE) {
             continue;
         }
-        if (issue.regionLayer == REGION_LAYER_VOXELS) {
-            addRegionsTasks(issue.regionLayer, ConvertTaskType::UPGRADE_VOXELS);
-        } else {
-            addRegionsTasks(issue.regionLayer, ConvertTaskType::UPGRADE_REGION);
-        }
+        addRegionsTasks(issue.regionLayer, ConvertTaskType::UPGRADE_REGION);
     }
 }
 
@@ -159,12 +156,13 @@ std::shared_ptr<Task> WorldConverter::startTask(
     return pool;
 }
 
-void WorldConverter::upgradeRegion(const fs::path& file, int x, int z) const {
-    throw std::runtime_error("unsupported region format");
-}
-
-void WorldConverter::upgradeVoxels(const fs::path& file, int x, int z) const {
-    throw std::runtime_error("unsupported region format");
+void WorldConverter::upgradeRegion(
+    const fs::path& file, int x, int z, RegionLayerIndex layer
+) const {
+    auto path = wfile->getRegions().getRegionFilePath(layer, x, z);
+    auto bytes = files::read_bytes_buffer(path);
+    auto buffer = compatibility::convertRegion2to3(bytes, layer);
+    files::write_bytes(path, buffer.data(), buffer.size());
 }
 
 void WorldConverter::convertVoxels(const fs::path& file, int x, int z) const {
@@ -195,11 +193,7 @@ void WorldConverter::convert(const ConvertTask& task) const {
 
     switch (task.type) {
         case ConvertTaskType::UPGRADE_REGION:
-            upgradeRegion(task.file, task.x, task.z);
-            break;
-        case ConvertTaskType::UPGRADE_VOXELS:
-            upgradeRegion(task.file, task.x, task.z);
-            upgradeVoxels(task.file, task.x, task.z);
+            upgradeRegion(task.file, task.x, task.z, task.layer);
             break;
         case ConvertTaskType::VOXELS:
             convertVoxels(task.file, task.x, task.z);

src/files/WorldConverter.hpp

@@ -24,8 +24,6 @@ enum class ConvertTaskType {
     PLAYER,
     /// @brief refresh region file version
     UPGRADE_REGION,
-    /// @brief rewrite voxels region file to new format
-    UPGRADE_VOXELS,
 };
 
 struct ConvertTask {
@@ -34,6 +32,7 @@ struct ConvertTask {
 
     /// @brief region coords
     int x, z;
+    RegionLayerIndex layer;
 };
 
 class WorldConverter : public Task {
@@ -45,8 +44,8 @@ class WorldConverter : public Task {
     uint tasksDone = 0;
     bool upgradeMode;
 
-    void upgradeRegion(const fs::path& file, int x, int z) const;
-    void upgradeVoxels(const fs::path& file, int x, int z) const;
+    void upgradeRegion(
+        const fs::path& file, int x, int z, RegionLayerIndex layer) const;
     void convertPlayer(const fs::path& file) const;
     void convertVoxels(const fs::path& file, int x, int z) const;
     void convertInventories(const fs::path& file, int x, int z) const;

src/files/WorldRegions.cpp

@@ -290,6 +290,10 @@ const fs::path& WorldRegions::getRegionsFolder(RegionLayerIndex layerid) const {
     return layers[layerid].folder;
 }
 
+fs::path WorldRegions::getRegionFilePath(RegionLayerIndex layerid, int x, int z) const {
+    return layers[layerid].getRegionFilePath(x, z);
+}
+
 void WorldRegions::writeAll() {
     for (auto& layer : layers) {
         fs::create_directories(layer.folder);

src/files/WorldRegions.hpp

@@ -147,6 +147,8 @@ struct RegionsLayer {
     WorldRegion* getRegion(int x, int z);
     WorldRegion* getOrCreateRegion(int x, int z);
 
+    fs::path getRegionFilePath(int x, int z) const;
+
     /// @brief Get chunk data. Read from file if not loaded yet.
     /// @param x chunk x coord
     /// @param z chunk z coord
@@ -237,6 +239,8 @@ public:
     /// @return directory path
     const fs::path& getRegionsFolder(RegionLayerIndex layerid) const;
 
+    fs::path getRegionFilePath(RegionLayerIndex layerid, int x, int z) const;
+
     /// @brief Write all region layers
     void writeAll();
 

src/files/compatibility.cpp

@@ -0,0 +1,109 @@
+#include "compatibility.hpp"
+
+#include <stdexcept>
+
+#include "constants.hpp"
+#include "voxels/voxel.hpp"
+#include "coders/compression.hpp"
+#include "coders/byte_utils.hpp"
+#include "lighting/Lightmap.hpp"
+#include "util/data_io.hpp"
+
+static inline size_t VOXELS_DATA_SIZE_V1 = CHUNK_VOL * 4;
+static inline size_t VOXELS_DATA_SIZE_V2 = CHUNK_VOL * 4;
+
+static util::Buffer<ubyte> convert_voxels_1to2(const ubyte* buffer, uint32_t size) {
+    auto data = compression::decompress(
+        buffer, size, VOXELS_DATA_SIZE_V1, compression::Method::EXTRLE8);
+
+    util::Buffer<ubyte> dstBuffer(VOXELS_DATA_SIZE_V2);
+    auto dst = reinterpret_cast<uint16_t*>(dstBuffer.data());
+
+    for (size_t i = 0; i < CHUNK_VOL; i++) {
+        ubyte bid1 = data[i];
+        ubyte bid2 = data[CHUNK_VOL + i];
+
+        ubyte bst1 = data[CHUNK_VOL * 2 + i];
+        ubyte bst2 = data[CHUNK_VOL * 3 + i];
+
+        dst[i] =
+            (static_cast<blockid_t>(bid1) << 8) | static_cast<blockid_t>(bid2);
+        dst[CHUNK_VOL + i] = (
+            (static_cast<blockstate_t>(bst1) << 8) |
+            static_cast<blockstate_t>(bst2)
+        );
+    }
+    size_t outLen;
+    auto compressed = compression::compress(
+        data.get(), VOXELS_DATA_SIZE_V2, outLen, compression::Method::EXTRLE16);
+    return util::Buffer<ubyte>(std::move(compressed), outLen);
+}
+
+util::Buffer<ubyte> compatibility::convertRegion2to3(
+    const util::Buffer<ubyte>& src, RegionLayerIndex layer
+) {
+    const size_t REGION_CHUNKS = 1024;
+    const size_t HEADER_SIZE = 10;
+    const size_t OFFSET_TABLE_SIZE = REGION_CHUNKS * sizeof(uint32_t);
+    const ubyte COMPRESS_NONE = 0;
+    const ubyte COMPRESS_EXTRLE8 = 1;
+    const ubyte COMPRESS_EXTRLE16 = 2;
+
+    const ubyte* const ptr = src.data();
+
+    ByteBuilder builder;
+    builder.putCStr(".VOXREG");
+    builder.put(3);
+    switch (layer) {
+        case REGION_LAYER_VOXELS: builder.put(COMPRESS_EXTRLE16); break;
+        case REGION_LAYER_LIGHTS: builder.put(COMPRESS_EXTRLE8); break;
+        default: builder.put(COMPRESS_NONE); break;
+    }
+
+    uint32_t offsets[REGION_CHUNKS] {};
+    size_t chunkIndex = 0;
+
+    auto tablePtr = reinterpret_cast<const uint32_t*>(
+        ptr + src.size() - OFFSET_TABLE_SIZE
+    );
+
+    for (size_t i = 0; i < REGION_CHUNKS; i++) {
+        uint32_t srcOffset = dataio::be2h(tablePtr[i]);
+        if (srcOffset == 0) {
+            continue;
+        }
+        uint32_t size = *reinterpret_cast<const uint32_t*>(ptr + srcOffset);
+        size = dataio::be2h(size);
+
+        const ubyte* data = ptr + srcOffset + sizeof(uint32_t);
+        offsets[i] = builder.size();
+
+        switch (layer) {
+            case REGION_LAYER_VOXELS: {
+                auto dstdata = convert_voxels_1to2(data, size);
+                builder.putInt32(dstdata.size());
+                builder.putInt32(VOXELS_DATA_SIZE_V2);
+                builder.put(dstdata.data(), dstdata.size());
+                break;
+            }
+            case REGION_LAYER_LIGHTS:
+                builder.putInt32(size);
+                builder.putInt32(LIGHTMAP_DATA_LEN);
+                builder.put(data, size);
+                break;
+            case REGION_LAYER_ENTITIES:
+            case REGION_LAYER_INVENTORIES: {
+                builder.putInt32(size);
+                builder.putInt32(size);
+                builder.put(data, size);
+                break;
+            case REGION_LAYERS_COUNT: 
+                throw std::invalid_argument("invalid enum");
+            }
+        }
+    }
+    for (size_t i = 0; i < REGION_CHUNKS; i++) {
+        builder.putInt32(offsets[i]);
+    }
+    return util::Buffer<ubyte>(builder.build().data(), builder.size());
+}

src/files/compatibility.hpp

@@ -0,0 +1,14 @@
+#pragma once
+
+#include "typedefs.hpp"
+#include "util/Buffer.hpp"
+#include "files/world_regions_fwd.hpp"
+
+namespace compatibility {
+    /// @brief Convert region file from version 2 to 3
+    /// @see /doc/specs/region_file_spec.md
+    /// @param src region file source content
+    /// @return new region file content
+    util::Buffer<ubyte> convertRegion2to3(
+        const util::Buffer<ubyte>& src, RegionLayerIndex layer);
+}

src/files/files.cpp

@@ -66,6 +66,12 @@ bool files::read(const fs::path& filename, char* data, size_t size) {
     return true;
 }
 
+util::Buffer<ubyte> files::read_bytes_buffer(const fs::path& path) {
+    size_t size;
+    auto bytes = files::read_bytes(path, size);
+    return util::Buffer<ubyte>(std::move(bytes), size);
+}
+
 std::unique_ptr<ubyte[]> files::read_bytes(
     const fs::path& filename, size_t& length
 ) {

src/files/files.hpp

@@ -7,6 +7,7 @@
 #include <vector>
 
 #include "typedefs.hpp"
+#include "util/Buffer.hpp"
 
 namespace fs = std::filesystem;
 
@@ -59,6 +60,7 @@ namespace files {
     );
 
     bool read(const fs::path&, char* data, size_t size);
+    util::Buffer<ubyte> read_bytes_buffer(const fs::path&);
     std::unique_ptr<ubyte[]> read_bytes(const fs::path&, size_t& length);
     std::vector<ubyte> read_bytes(const fs::path&);
     std::string read_string(const fs::path& filename);

test/files/compatibility.cpp

@@ -0,0 +1,15 @@
+#include <gtest/gtest.h>
+#include <filesystem>
+
+#include "files/files.hpp"
+#include "files/compatibility.hpp"
+
+TEST(compatibility, convert) {
+    auto infile = std::filesystem::u8path(
+        "voxels_0_1.bin");
+    auto outfile = std::filesystem::u8path(
+        "output_0_1.bin");
+    auto input = files::read_bytes_buffer(infile);
+    auto output = compatibility::convertRegion2to3(input, REGION_LAYER_VOXELS);
+    files::write_bytes(outfile, output.data(), output.size());
+}