1 files changed, 964 insertions, 0 deletions

diff --git a/source/WorldStorage/WSSCompact.cpp b/source/WorldStorage/WSSCompact.cpp
new file mode 100644
index 000000000..e604ca9b9
--- /dev/null
+++ b/source/WorldStorage/WSSCompact.cpp
@@ -0,0 +1,964 @@
+
+// WSSCompact.cpp
+
+// Interfaces to the cWSSCompact class representing the "compact" storage schema (PAK-files)
+
+#include "Globals.h"
+#include "WSSCompact.h"
+#include "../cWorld.h"
+#include "zlib.h"
+#include <json/json.h>
+#include "../StringCompression.h"
+#include "../cChestEntity.h"
+#include "../cSignEntity.h"
+#include "../cFurnaceEntity.h"
+#include "../cNoteEntity.h"
+#include "../BlockID.h"
+
+
+
+
+
+#pragma pack(push, 1)
+/// The chunk header, as stored in the file:
+struct cWSSCompact::sChunkHeader
+{
+	int m_ChunkX;
+	int m_ChunkZ;
+	int m_CompressedSize;
+	int m_UncompressedSize;
+} ;
+#pragma pack(pop)
+
+
+
+
+
+/// The maximum number of PAK files that are cached
+const int MAX_PAK_FILES = 16;
+
+/// The maximum number of unsaved chunks before the cPAKFile saves them to disk
+const int MAX_DIRTY_CHUNKS = 16;
+
+
+
+
+
+///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+// cJsonChunkSerializer:
+
+cJsonChunkSerializer::cJsonChunkSerializer(void) :
+	m_HasJsonData(false)
+{
+}
+
+
+
+
+
+void cJsonChunkSerializer::Entity(cEntity * a_Entity)
+{
+	// TODO: a_Entity->SaveToJson(m_Root);
+}
+
+
+
+
+
+void cJsonChunkSerializer::BlockEntity(cBlockEntity * a_BlockEntity)
+{
+	const char * SaveInto = NULL;
+	switch (a_BlockEntity->GetBlockType())
+	{
+		case E_BLOCK_CHEST:     SaveInto = "Chests";   break;
+		case E_BLOCK_FURNACE:   SaveInto = "Furnaces"; break;
+		case E_BLOCK_SIGN_POST: SaveInto = "Signs";    break;
+		case E_BLOCK_WALLSIGN:  SaveInto = "Signs";    break;
+		case E_BLOCK_NOTE_BLOCK:  SaveInto = "Notes";    break;
+
+		default:
+		{
+			ASSERT(!"Unhandled blocktype in BlockEntities list while saving to JSON");
+			break;
+		}
+	}  // switch (BlockEntity->GetBlockType())
+	if (SaveInto == NULL)
+	{
+		return;
+	}
+	
+	Json::Value val;
+	a_BlockEntity->SaveToJson(val);
+	m_Root[SaveInto].append(val);
+	m_HasJsonData = true;
+}
+
+
+
+
+
+bool cJsonChunkSerializer::LightIsValid(bool a_IsLightValid)
+{
+	if (!a_IsLightValid)
+	{
+		return false;
+	}
+	m_Root["IsLightValid"] = true;
+	m_HasJsonData = true;
+	return true;
+}
+
+
+
+
+
+///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+// cWSSCompact:
+
+cWSSCompact::~cWSSCompact()
+{
+	for (cPAKFiles::iterator itr = m_PAKFiles.begin(); itr != m_PAKFiles.end(); ++itr)
+	{
+		delete *itr;
+	}
+}
+
+
+
+
+
+bool cWSSCompact::LoadChunk(const cChunkCoords & a_Chunk)
+{
+	AString ChunkData;
+	int UncompressedSize = 0;
+	if (!GetChunkData(a_Chunk, UncompressedSize, ChunkData))
+	{
+		// The reason for failure is already printed in GetChunkData()
+		return false;
+	}
+	
+	return LoadChunkFromData(a_Chunk, UncompressedSize, ChunkData, m_World);
+}
+
+
+
+
+
+bool cWSSCompact::SaveChunk(const cChunkCoords & a_Chunk)
+{
+	cCSLock Lock(m_CS);
+	
+	cPAKFile * f = LoadPAKFile(a_Chunk);
+	if (f == NULL)
+	{
+		// For some reason we couldn't locate the file
+		LOG("Cannot locate a proper PAK file for chunk [%d, %d]", a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ);
+		return false;
+	}
+	return f->SaveChunk(a_Chunk, m_World);
+}
+
+
+
+
+
+cWSSCompact::cPAKFile * cWSSCompact::LoadPAKFile(const cChunkCoords & a_Chunk)
+{
+	// ASSUMES that m_CS has been locked
+	
+	// We need to retain this weird conversion code, because some edge chunks are in the wrong PAK file
+	const int LayerX = (int)(floorf((float)a_Chunk.m_ChunkX / 32.0f));
+	const int LayerZ = (int)(floorf((float)a_Chunk.m_ChunkZ / 32.0f));
+	
+	// Is it already cached?
+	for (cPAKFiles::iterator itr = m_PAKFiles.begin(); itr != m_PAKFiles.end(); ++itr)
+	{
+		if (((*itr) != NULL) && ((*itr)->GetLayerX() == LayerX) && ((*itr)->GetLayerZ() == LayerZ))
+		{
+			// Move the file to front and return it:
+			cPAKFile * f = *itr;
+			if (itr != m_PAKFiles.begin())
+			{
+				m_PAKFiles.erase(itr);
+				m_PAKFiles.push_front(f);
+			}
+			return f;
+		}
+	}
+	
+	// Load it anew:
+	AString FileName;
+	Printf(FileName, "%s/X%i_Z%i.pak", m_World->GetName().c_str(), LayerX, LayerZ );
+	cPAKFile * f = new cPAKFile(FileName, LayerX, LayerZ);
+	if (f == NULL)
+	{
+		return NULL;
+	}
+	m_PAKFiles.push_front(f);
+	
+	// If there are too many PAK files cached, delete the last one used:
+	if (m_PAKFiles.size() > MAX_PAK_FILES)
+	{
+		delete m_PAKFiles.back();
+		m_PAKFiles.pop_back();
+	}
+	return f;
+}
+
+
+
+
+
+bool cWSSCompact::GetChunkData(const cChunkCoords & a_Chunk, int & a_UncompressedSize, AString & a_Data)
+{
+	cCSLock Lock(m_CS);
+	cPAKFile * f = LoadPAKFile(a_Chunk);
+	if (f == NULL)
+	{
+		return false;
+	}
+	return f->GetChunkData(a_Chunk, a_UncompressedSize, a_Data);
+}
+
+
+
+
+
+/*
+// TODO: Rewrite saving to use the same principles as loading
+bool cWSSCompact::SetChunkData(const cChunkCoords & a_Chunk, int a_UncompressedSize, const AString & a_Data)
+{
+	cCSLock Lock(m_CS);
+	cPAKFile * f = LoadPAKFile(a_Chunk);
+	if (f == NULL)
+	{
+		return false;
+	}
+	return f->SetChunkData(a_Chunk, a_UncompressedSize, a_Data);
+}
+*/
+
+
+
+
+
+bool cWSSCompact::EraseChunkData(const cChunkCoords & a_Chunk)
+{
+	cCSLock Lock(m_CS);
+	cPAKFile * f = LoadPAKFile(a_Chunk);
+	if (f == NULL)
+	{
+		return false;
+	}
+	return f->EraseChunkData(a_Chunk);
+}
+
+
+
+
+
+void cWSSCompact::LoadEntitiesFromJson(Json::Value & a_Value, cEntityList & a_Entities, cBlockEntityList & a_BlockEntities, cWorld * a_World)
+{
+	// Load chests
+	Json::Value AllChests = a_Value.get("Chests", Json::nullValue);
+	if (!AllChests.empty())
+	{
+		for (Json::Value::iterator itr = AllChests.begin(); itr != AllChests.end(); ++itr )
+		{
+			Json::Value & Chest = *itr;
+			cChestEntity * ChestEntity = new cChestEntity(0,0,0, a_World);
+			if (!ChestEntity->LoadFromJson( Chest ) )
+			{
+				LOGERROR("ERROR READING CHEST FROM JSON!" );
+				delete ChestEntity;
+			}
+			else
+			{
+				a_BlockEntities.push_back( ChestEntity );
+			}
+		}  // for itr - AllChests[]
+	}
+
+	// Load furnaces
+	Json::Value AllFurnaces = a_Value.get("Furnaces", Json::nullValue);
+	if( !AllFurnaces.empty() )
+	{
+		for( Json::Value::iterator itr = AllFurnaces.begin(); itr != AllFurnaces.end(); ++itr )
+		{
+			Json::Value & Furnace = *itr;
+			cFurnaceEntity * FurnaceEntity = new cFurnaceEntity(0,0,0, a_World);
+			if( !FurnaceEntity->LoadFromJson( Furnace ) )
+			{
+				LOGERROR("ERROR READING FURNACE FROM JSON!" );
+				delete FurnaceEntity;
+			}
+			else
+			{
+				a_BlockEntities.push_back( FurnaceEntity );
+			}
+		}  // for itr - AllFurnaces[]
+	}
+
+	// Load signs
+	Json::Value AllSigns = a_Value.get("Signs", Json::nullValue);
+	if( !AllSigns.empty() )
+	{
+		for( Json::Value::iterator itr = AllSigns.begin(); itr != AllSigns.end(); ++itr )
+		{
+			Json::Value & Sign = *itr;
+			cSignEntity * SignEntity = new cSignEntity( E_BLOCK_SIGN_POST, 0,0,0, a_World);
+			if ( !SignEntity->LoadFromJson( Sign ) )
+			{
+				LOGERROR("ERROR READING SIGN FROM JSON!" );
+				delete SignEntity;
+			}
+			else
+			{
+				a_BlockEntities.push_back( SignEntity );
+			}
+		}  // for itr - AllSigns[]
+	}
+
+	// Load note blocks
+	Json::Value AllNotes = a_Value.get("Notes", Json::nullValue);
+	if( !AllNotes.empty() )
+	{
+		for( Json::Value::iterator itr = AllNotes.begin(); itr != AllNotes.end(); ++itr )
+		{
+			Json::Value & Note = *itr;
+			cNoteEntity * NoteEntity = new cNoteEntity(0, 0, 0, a_World);
+			if ( !NoteEntity->LoadFromJson( Note ) )
+			{
+				LOGERROR("ERROR READING NOTE BLOCK FROM JSON!" );
+				delete NoteEntity;
+			}
+			else
+			{
+				a_BlockEntities.push_back( NoteEntity );
+			}
+		}  // for itr - AllNotes[]
+	}
+}
+
+
+
+
+
+///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
+// cWSSCompact::cPAKFile
+
+#define READ(Var) \
+	if (f.Read(&Var, sizeof(Var)) != sizeof(Var)) \
+	{ \
+		LOGERROR("ERROR READING %s FROM FILE %s (line %d); file offset %d", #Var, m_FileName.c_str(), __LINE__, f.Tell()); \
+		return; \
+	}
+
+cWSSCompact::cPAKFile::cPAKFile(const AString & a_FileName, int a_LayerX, int a_LayerZ) :
+	m_FileName(a_FileName),
+	m_LayerX(a_LayerX),
+	m_LayerZ(a_LayerZ),
+	m_NumDirty(0),
+	m_ChunkVersion( CHUNK_VERSION ), // Init with latest version
+	m_PakVersion( PAK_VERSION )
+{
+	cFile f;
+	if (!f.Open(m_FileName, cFile::fmRead))
+	{
+		return;
+	}
+	
+	// Read headers:
+	READ(m_PakVersion);
+	if (m_PakVersion != 1)
+	{
+		LOGERROR("File \"%s\" is in an unknown pak format (%d)", m_FileName.c_str(), m_PakVersion);
+		return;
+	}
+	
+	READ(m_ChunkVersion);
+	switch( m_ChunkVersion )
+	{
+	case 1:
+		m_ChunkSize.Set(16, 128, 16);
+		break;
+	case 2:
+	case 3:
+		m_ChunkSize.Set(16, 256, 16);
+		break;
+	default:
+		LOGERROR("File \"%s\" is in an unknown chunk format (%d)", m_FileName.c_str(), m_ChunkVersion);
+		return;
+	};
+	
+	short NumChunks = 0;
+	READ(NumChunks);
+	
+	// Read chunk headers:
+	for (int i = 0; i < NumChunks; i++)
+	{
+		sChunkHeader * Header = new sChunkHeader;
+		READ(*Header);
+		m_ChunkHeaders.push_back(Header);
+	}  // for i - chunk headers
+
+	// Read chunk data:
+	if (f.ReadRestOfFile(m_DataContents) == -1)
+	{
+		LOGERROR("Cannot read file \"%s\" contents", m_FileName.c_str());
+		return;
+	}
+
+	if( m_ChunkVersion == 1 ) // Convert chunks to version 2
+	{
+		UpdateChunk1To2();
+	}
+#if AXIS_ORDER == AXIS_ORDER_XZY
+	if( m_ChunkVersion == 2 ) // Convert chunks to version 3
+	{
+		UpdateChunk2To3();
+	}
+#endif
+}
+
+
+
+
+
+cWSSCompact::cPAKFile::~cPAKFile()
+{
+	if (m_NumDirty > 0)
+	{
+		SynchronizeFile();
+	}
+	for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+	{
+		delete *itr;
+	}
+}
+
+
+
+
+
+bool cWSSCompact::cPAKFile::GetChunkData(const cChunkCoords & a_Chunk, int & a_UncompressedSize, AString & a_Data)
+{
+	int ChunkX = a_Chunk.m_ChunkX;
+	int ChunkZ = a_Chunk.m_ChunkZ;
+	sChunkHeader * Header = NULL;
+	int Offset = 0;
+	for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+	{
+		if (((*itr)->m_ChunkX == ChunkX) && ((*itr)->m_ChunkZ == ChunkZ))
+		{
+			Header = *itr;
+			break;
+		}
+		Offset += (*itr)->m_CompressedSize;
+	}
+	if ((Header == NULL) || (Offset + Header->m_CompressedSize > (int)m_DataContents.size()))
+	{
+		// Chunk not found / data invalid
+		return false;
+	}
+
+	a_UncompressedSize = Header->m_UncompressedSize;
+	a_Data.assign(m_DataContents, Offset, Header->m_CompressedSize);
+	return true;
+}
+
+
+
+
+
+bool cWSSCompact::cPAKFile::SaveChunk(const cChunkCoords & a_Chunk, cWorld * a_World)
+{
+	if (!SaveChunkToData(a_Chunk, a_World))
+	{
+		return false;
+	}
+	if (m_NumDirty > MAX_DIRTY_CHUNKS)
+	{
+		SynchronizeFile();
+	}
+	return true;
+}
+
+
+
+
+
+void cWSSCompact::cPAKFile::UpdateChunk1To2()
+{
+	int Offset = 0;
+	AString NewDataContents;
+	int ChunksConverted = 0;
+	for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+	{
+		sChunkHeader * Header = *itr;
+
+		if( ChunksConverted % 32 == 0 )
+		{
+			LOGINFO("Updating \"%s\" version 1 to version 2: %d %%", m_FileName.c_str(), (ChunksConverted * 100) / m_ChunkHeaders.size() );
+		}
+		ChunksConverted++;
+
+		AString Data;
+		int UncompressedSize = Header->m_UncompressedSize;
+		Data.assign(m_DataContents, Offset, Header->m_CompressedSize);
+		Offset += Header->m_CompressedSize;
+
+		// Crude data integrity check:
+		int ExpectedSize = (16*128*16)*2 + (16*128*16)/2; // For version 1
+		if (UncompressedSize < ExpectedSize)
+		{
+			LOGWARNING("Chunk [%d, %d] has too short decompressed data (%d bytes out of %d needed), erasing",
+				Header->m_ChunkX, Header->m_ChunkZ,
+				UncompressedSize, ExpectedSize
+				);
+			Offset += Header->m_CompressedSize;
+			continue;
+		}
+
+		// Decompress the data:
+		AString UncompressedData;
+		{
+			int errorcode = UncompressString(Data.data(), Data.size(), UncompressedData, UncompressedSize);
+			if (errorcode != Z_OK)
+			{
+				LOGERROR("Error %d decompressing data for chunk [%d, %d]", 
+					errorcode,
+					Header->m_ChunkX, Header->m_ChunkZ
+					);
+				Offset += Header->m_CompressedSize;
+				continue;
+			}
+		}
+
+		if (UncompressedSize != (int)UncompressedData.size())
+		{
+			LOGWARNING("Uncompressed data size differs (exp %d bytes, got %d) for chunk [%d, %d]",
+				UncompressedSize, UncompressedData.size(),
+				Header->m_ChunkX, Header->m_ChunkZ
+				);
+			Offset += Header->m_CompressedSize;
+			continue;
+		}
+
+
+		// Old version is 128 blocks high with YZX axis order
+		char ConvertedData[cChunkDef::BlockDataSize];
+		int Index = 0;
+		unsigned int InChunkOffset = 0;
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) 
+		{
+			for( int y = 0; y < 128; ++y )
+			{
+				ConvertedData[Index++] = UncompressedData[y + z * 128 + x * 128 * 16 + InChunkOffset];
+			}
+			// Add 128 empty blocks after an old y column
+			memset(ConvertedData + Index, E_BLOCK_AIR, 128);
+			Index += 128;
+		}
+		InChunkOffset += (16 * 128 * 16);
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) // Metadata
+		{
+			for( int y = 0; y < 64; ++y )
+			{
+				ConvertedData[Index++] = UncompressedData[y + z * 64 + x * 64 * 16 + InChunkOffset];
+			}
+			memset(ConvertedData + Index, 0, 64);
+			Index += 64;
+		}
+		InChunkOffset += (16 * 128 * 16) / 2;
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) // Block light
+		{
+			for( int y = 0; y < 64; ++y )
+			{
+				ConvertedData[Index++] = UncompressedData[y + z * 64 + x * 64 * 16 + InChunkOffset];
+			}
+			memset(ConvertedData + Index, 0, 64);
+			Index += 64;
+		}
+		InChunkOffset += (16*128*16)/2;
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) // Sky light
+		{
+			for( int y = 0; y < 64; ++y )
+			{
+				ConvertedData[Index++] = UncompressedData[y + z * 64 + x * 64 * 16 + InChunkOffset];
+			}
+			memset(ConvertedData + Index, 0, 64);
+			Index += 64;
+		}
+		InChunkOffset += (16 * 128 * 16) / 2;
+
+		AString Converted(ConvertedData, ARRAYCOUNT(ConvertedData));
+		
+		// Add JSON data afterwards
+		if (UncompressedData.size() > InChunkOffset)
+		{
+			Converted.append( UncompressedData.begin() + InChunkOffset, UncompressedData.end() );
+		}
+
+		// Re-compress data
+		AString CompressedData;
+		{
+			int errorcode = CompressString(Converted.data(), Converted.size(), CompressedData);
+			if (errorcode != Z_OK)
+			{
+				LOGERROR("Error %d compressing data for chunk [%d, %d]", 
+					errorcode,
+					Header->m_ChunkX, Header->m_ChunkZ
+				);
+				continue;
+			}
+		}
+
+		// Save into file's cache
+		Header->m_UncompressedSize = Converted.size();
+		Header->m_CompressedSize = CompressedData.size();
+		NewDataContents.append( CompressedData );
+	}
+
+	// Done converting
+	m_DataContents = NewDataContents;
+	m_ChunkVersion = 2;
+	SynchronizeFile();
+	
+	LOGINFO("Updated \"%s\" version 1 to version 2", m_FileName.c_str() );
+}
+
+
+
+
+
+void cWSSCompact::cPAKFile::UpdateChunk2To3()
+{
+	int Offset = 0;
+	AString NewDataContents;
+	int ChunksConverted = 0;
+	for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+	{
+		sChunkHeader * Header = *itr;
+
+		if( ChunksConverted % 32 == 0 )
+		{
+			LOGINFO("Updating \"%s\" version 2 to version 3: %d %%", m_FileName.c_str(), (ChunksConverted * 100) / m_ChunkHeaders.size() );
+		}
+		ChunksConverted++;
+
+		AString Data;
+		int UncompressedSize = Header->m_UncompressedSize;
+		Data.assign(m_DataContents, Offset, Header->m_CompressedSize);
+		Offset += Header->m_CompressedSize;
+
+		// Crude data integrity check:
+		const int ExpectedSize = (16*256*16)*2 + (16*256*16)/2; // For version 2
+		if (UncompressedSize < ExpectedSize)
+		{
+			LOGWARNING("Chunk [%d, %d] has too short decompressed data (%d bytes out of %d needed), erasing",
+				Header->m_ChunkX, Header->m_ChunkZ,
+				UncompressedSize, ExpectedSize
+				);
+			Offset += Header->m_CompressedSize;
+			continue;
+		}
+
+		// Decompress the data:
+		AString UncompressedData;
+		{
+			int errorcode = UncompressString(Data.data(), Data.size(), UncompressedData, UncompressedSize);
+			if (errorcode != Z_OK)
+			{
+				LOGERROR("Error %d decompressing data for chunk [%d, %d]", 
+					errorcode,
+					Header->m_ChunkX, Header->m_ChunkZ
+					);
+				Offset += Header->m_CompressedSize;
+				continue;
+			}
+		}
+
+		if (UncompressedSize != (int)UncompressedData.size())
+		{
+			LOGWARNING("Uncompressed data size differs (exp %d bytes, got %d) for chunk [%d, %d]",
+				UncompressedSize, UncompressedData.size(),
+				Header->m_ChunkX, Header->m_ChunkZ
+				);
+			Offset += Header->m_CompressedSize;
+			continue;
+		}
+
+		char ConvertedData[ExpectedSize];
+		memset(ConvertedData, 0, ExpectedSize);
+
+		// Cannot use cChunk::MakeIndex because it might change again?????????
+		// For compatibility, use what we know is current
+		#define MAKE_2_INDEX( x, y, z ) ( y + (z * 256) + (x * 256 * 16) )
+		#define MAKE_3_INDEX( x, y, z ) ( x + (z * 16) + (y * 16 * 16) )
+
+		unsigned int InChunkOffset = 0;
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) for( int y = 0; y < 256; ++y ) // YZX Loop order is important, in 1.1 Y was first then Z then X
+		{
+			ConvertedData[ MAKE_3_INDEX(x, y, z) ] = UncompressedData[InChunkOffset];
+			++InChunkOffset;
+		}  // for y, z, x
+
+		
+		unsigned int index2 = 0;
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) for( int y = 0; y < 256; ++y )
+		{
+			ConvertedData[ InChunkOffset + MAKE_3_INDEX(x, y, z)/2 ] |= ( (UncompressedData[ InChunkOffset + index2/2 ] >> ((index2&1)*4) ) & 0x0f ) << ((x&1)*4);
+			++index2;
+		}
+		InChunkOffset += index2 / 2;
+		index2 = 0;
+
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) for( int y = 0; y < 256; ++y )
+		{
+			ConvertedData[ InChunkOffset + MAKE_3_INDEX(x, y, z)/2 ] |= ( (UncompressedData[ InChunkOffset + index2/2 ] >> ((index2&1)*4) ) & 0x0f ) << ((x&1)*4);
+			++index2;
+		}
+		InChunkOffset += index2 / 2;
+		index2 = 0;
+
+		for( int x = 0; x < 16; ++x ) for( int z = 0; z < 16; ++z ) for( int y = 0; y < 256; ++y )
+		{
+			ConvertedData[ InChunkOffset + MAKE_3_INDEX(x, y, z)/2 ] |= ( (UncompressedData[ InChunkOffset + index2/2 ] >> ((index2&1)*4) ) & 0x0f ) << ((x&1)*4);
+			++index2;
+		}
+		InChunkOffset += index2 / 2;
+		index2 = 0;
+
+		AString Converted(ConvertedData, ExpectedSize);
+
+		// Add JSON data afterwards
+		if (UncompressedData.size() > InChunkOffset)
+		{
+			Converted.append( UncompressedData.begin() + InChunkOffset, UncompressedData.end() );
+		}
+
+		// Re-compress data
+		AString CompressedData;
+		{
+			int errorcode = CompressString(Converted.data(), Converted.size(), CompressedData);
+			if (errorcode != Z_OK)
+			{
+				LOGERROR("Error %d compressing data for chunk [%d, %d]", 
+					errorcode,
+					Header->m_ChunkX, Header->m_ChunkZ
+					);
+				continue;
+			}
+		}
+
+		// Save into file's cache
+		Header->m_UncompressedSize = Converted.size();
+		Header->m_CompressedSize = CompressedData.size();
+		NewDataContents.append( CompressedData );
+	}
+
+	// Done converting
+	m_DataContents = NewDataContents;
+	m_ChunkVersion = 3;
+	SynchronizeFile();
+
+	LOGINFO("Updated \"%s\" version 2 to version 3", m_FileName.c_str() );
+}
+
+
+
+
+
+bool cWSSCompact::LoadChunkFromData(const cChunkCoords & a_Chunk, int & a_UncompressedSize, const AString & a_Data, cWorld * a_World)
+{
+	// Crude data integrity check:
+	if (a_UncompressedSize < cChunkDef::BlockDataSize)
+	{
+		LOGWARNING("Chunk [%d, %d] has too short decompressed data (%d bytes out of %d needed), erasing",
+			a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ,
+			a_UncompressedSize, cChunkDef::BlockDataSize
+		);
+		EraseChunkData(a_Chunk);
+		return false;
+	}
+	
+	// Decompress the data:
+	AString UncompressedData;
+	int errorcode = UncompressString(a_Data.data(), a_Data.size(), UncompressedData, a_UncompressedSize);
+	if (errorcode != Z_OK)
+	{
+		LOGERROR("Error %d decompressing data for chunk [%d, %d]", 
+			errorcode,
+			a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ
+		);
+		return false;
+	}
+	
+	if (a_UncompressedSize != (int)UncompressedData.size())
+	{
+		LOGWARNING("Uncompressed data size differs (exp %d bytes, got %d) for chunk [%d, %d]",
+			a_UncompressedSize, UncompressedData.size(),
+			a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ
+		);
+		return false;
+	}
+
+	cEntityList      Entities;
+	cBlockEntityList BlockEntities;
+	bool IsLightValid = false;
+	
+	if (a_UncompressedSize > cChunkDef::BlockDataSize)
+	{
+		Json::Value root;   // will contain the root value after parsing.
+		Json::Reader reader;
+		if ( !reader.parse( UncompressedData.data() + cChunkDef::BlockDataSize, root, false ) )
+		{
+			LOGERROR("Failed to parse trailing JSON in chunk [%d, %d]!",
+				a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ
+			);
+		}
+		else
+		{
+			LoadEntitiesFromJson(root, Entities, BlockEntities, a_World);
+			IsLightValid = root.get("IsLightValid", false).asBool();
+		}
+	}
+
+	BLOCKTYPE *  BlockData  = (BLOCKTYPE *)UncompressedData.data();
+	NIBBLETYPE * MetaData   = (NIBBLETYPE *)(BlockData + cChunkDef::MetaOffset);
+	NIBBLETYPE * BlockLight = (NIBBLETYPE *)(BlockData + cChunkDef::LightOffset);
+	NIBBLETYPE * SkyLight   = (NIBBLETYPE *)(BlockData + cChunkDef::SkyLightOffset);
+	
+	a_World->SetChunkData(
+		a_Chunk.m_ChunkX, a_Chunk.m_ChunkY, a_Chunk.m_ChunkZ,
+		BlockData, MetaData,
+		IsLightValid ? BlockLight : NULL,
+		IsLightValid ? SkyLight : NULL,
+		NULL, NULL,
+		Entities, BlockEntities,
+		false
+	);
+
+	return true;
+}
+
+
+
+
+
+bool cWSSCompact::cPAKFile::EraseChunkData(const cChunkCoords & a_Chunk)
+{
+	int ChunkX = a_Chunk.m_ChunkX;
+	int ChunkZ = a_Chunk.m_ChunkZ;
+	int Offset = 0;
+	for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+	{
+		if (((*itr)->m_ChunkX == ChunkX) && ((*itr)->m_ChunkZ == ChunkZ))
+		{
+			m_DataContents.erase(Offset, (*itr)->m_CompressedSize);
+			delete *itr;
+			itr = m_ChunkHeaders.erase(itr);
+			return true;
+		}
+		Offset += (*itr)->m_CompressedSize;
+	}
+	
+	return false;
+}
+
+
+
+
+
+bool cWSSCompact::cPAKFile::SaveChunkToData(const cChunkCoords & a_Chunk, cWorld * a_World)
+{
+	// Serialize the chunk:
+	cJsonChunkSerializer Serializer;
+	if (!a_World->GetChunkData(a_Chunk.m_ChunkX, a_Chunk.m_ChunkY, a_Chunk.m_ChunkZ, Serializer))
+	{
+		// Chunk not valid
+		LOG("cWSSCompact: Trying to save chunk [%d, %d, %d] that has no data, ignoring request.", a_Chunk.m_ChunkX, a_Chunk.m_ChunkY, a_Chunk.m_ChunkZ);
+		return false;
+	}
+
+	AString Data;
+	Data.assign((const char *)Serializer.GetBlockData(), cChunkDef::BlockDataSize);
+	if (Serializer.HasJsonData())
+	{
+		AString JsonData;
+		Json::StyledWriter writer;
+		JsonData = writer.write(Serializer.GetRoot());
+		Data.append(JsonData);
+	}
+	
+	// Compress the data:
+	AString CompressedData;
+	int errorcode = CompressString(Data.data(), Data.size(), CompressedData);
+	if ( errorcode != Z_OK )
+	{
+		LOGERROR("Error %i compressing data for chunk [%d, %d, %d]", errorcode, a_Chunk.m_ChunkX, a_Chunk.m_ChunkY, a_Chunk.m_ChunkZ);
+		return false;
+	}
+	
+	// Erase any existing data for the chunk:
+	EraseChunkData(a_Chunk);
+	
+	// Save the header:
+	sChunkHeader * Header = new sChunkHeader;
+	if (Header == NULL)
+	{
+		LOGWARNING("Cannot create a new chunk header to save chunk [%d, %d, %d]", a_Chunk.m_ChunkX, a_Chunk.m_ChunkY, a_Chunk.m_ChunkZ);
+		return false;
+	}
+	Header->m_CompressedSize = (int)CompressedData.size();
+	Header->m_ChunkX = a_Chunk.m_ChunkX;
+	Header->m_ChunkZ = a_Chunk.m_ChunkZ;
+	Header->m_UncompressedSize = (int)Data.size();
+	m_ChunkHeaders.push_back(Header);
+	
+	m_DataContents.append(CompressedData.data(), CompressedData.size());
+	
+	m_NumDirty++;
+	return true;
+}
+
+
+
+
+
+#define WRITE(Var) \
+	if (f.Write(&Var, sizeof(Var)) != sizeof(Var)) \
+	{ \
+		LOGERROR("cWSSCompact: ERROR writing %s to file \"%s\" (line %d); file offset %d", #Var, m_FileName.c_str(), __LINE__, f.Tell()); \
+		return; \
+	}
+
+void cWSSCompact::cPAKFile::SynchronizeFile(void)
+{
+	cFile f;
+	if (!f.Open(m_FileName, cFile::fmWrite))
+	{
+		LOGERROR("Cannot open PAK file \"%s\" for writing", m_FileName.c_str());
+		return;
+	}
+	
+	WRITE(m_PakVersion);
+	WRITE(m_ChunkVersion);
+	short NumChunks = (short)m_ChunkHeaders.size();
+	WRITE(NumChunks);
+	for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+	{
+		WRITE(**itr);
+	}
+	if (f.Write(m_DataContents.data(), m_DataContents.size()) != (int)m_DataContents.size())
+	{
+		LOGERROR("cWSSCompact: ERROR writing chunk contents to file \"%s\" (line %d); file offset %d", m_FileName.c_str(), __LINE__, f.Tell());
+		return;
+	}
+	m_NumDirty = 0;
+}
+
+
+
+