2012-02-13 16:47:03 -05:00
|
|
|
|
|
|
|
// 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>
|
2012-02-16 08:42:35 -05:00
|
|
|
#include "StringCompression.h"
|
|
|
|
#include "cChestEntity.h"
|
|
|
|
#include "cSignEntity.h"
|
|
|
|
#include "cFurnaceEntity.h"
|
|
|
|
#include "BlockID.h"
|
2012-02-13 16:47:03 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
#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;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
// cWSSCompact:
|
|
|
|
|
|
|
|
cWSSCompact::~cWSSCompact()
|
|
|
|
{
|
|
|
|
for (cPAKFiles::iterator itr = m_PAKFiles.begin(); itr != m_PAKFiles.end(); ++itr)
|
|
|
|
{
|
|
|
|
delete *itr;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
bool cWSSCompact::LoadChunk(const cChunkCoords & a_Chunk)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
cPAKFile * f = LoadPAKFile(a_Chunk);
|
|
|
|
if (f == NULL)
|
|
|
|
{
|
|
|
|
// For some reason we couldn't locate the file
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
return f->LoadChunk(a_Chunk, m_World);
|
2012-02-13 16:47:03 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
bool cWSSCompact::SaveChunk(const cChunkCoords & a_Chunk)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
cPAKFile * f = LoadPAKFile(a_Chunk);
|
|
|
|
if (f == NULL)
|
|
|
|
{
|
|
|
|
// For some reason we couldn't locate the file
|
|
|
|
return false;
|
|
|
|
}
|
2012-02-16 08:42:35 -05:00
|
|
|
return f->SaveChunk(a_Chunk, m_World);
|
2012-02-13 16:47:03 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
cWSSCompact::cPAKFile * cWSSCompact::LoadPAKFile(const cChunkCoords & a_Chunk)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
// We need to retain this weird conversion code, because some edge chunks are in the wrong PAK file
|
2012-02-16 08:42:35 -05:00
|
|
|
const int LayerX = (int)(floorf((float)a_Chunk.m_ChunkX / 32.0f));
|
|
|
|
const int LayerZ = (int)(floorf((float)a_Chunk.m_ChunkZ / 32.0f));
|
2012-02-13 16:47:03 -05:00
|
|
|
|
|
|
|
// 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;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
///////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
// 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)
|
|
|
|
{
|
|
|
|
cFile f;
|
|
|
|
if (!f.Open(m_FileName, cFile::fmRead))
|
|
|
|
{
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Read headers:
|
|
|
|
char PakVersion = 0;
|
|
|
|
READ(PakVersion);
|
|
|
|
if (PakVersion != 1)
|
|
|
|
{
|
|
|
|
LOGERROR("File \"%s\" is in an unknown pak format (%d)", m_FileName.c_str(), PakVersion);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
char ChunkVersion = 0;
|
|
|
|
READ(ChunkVersion);
|
|
|
|
if (ChunkVersion != 1)
|
|
|
|
{
|
|
|
|
LOGERROR("File \"%s\" is in an unknown chunk format (%d)", m_FileName.c_str(), 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;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
cWSSCompact::cPAKFile::~cPAKFile()
|
|
|
|
{
|
|
|
|
if (m_NumDirty > 0)
|
|
|
|
{
|
|
|
|
SynchronizeFile();
|
|
|
|
}
|
|
|
|
for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
|
|
|
|
{
|
|
|
|
delete *itr;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
bool cWSSCompact::cPAKFile::LoadChunk(const cChunkCoords & a_Chunk, cWorld * a_World)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
2012-02-16 08:42:35 -05:00
|
|
|
int ChunkX = a_Chunk.m_ChunkX;
|
|
|
|
int ChunkZ = a_Chunk.m_ChunkZ;
|
2012-02-13 16:47:03 -05:00
|
|
|
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;
|
|
|
|
}
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
return LoadChunk(a_Chunk, Offset, Header, a_World);
|
2012-02-13 16:47:03 -05:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
bool cWSSCompact::cPAKFile::SaveChunk(const cChunkCoords & a_Chunk, cWorld * a_World)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
2012-02-16 08:42:35 -05:00
|
|
|
if (!SaveChunkToData(a_Chunk, a_World))
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
if (m_NumDirty > MAX_DIRTY_CHUNKS)
|
|
|
|
{
|
|
|
|
SynchronizeFile();
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
bool cWSSCompact::cPAKFile::LoadChunk(const cChunkCoords & a_Chunk, int a_Offset, sChunkHeader * a_Header, cWorld * a_World)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
2012-02-16 08:42:35 -05:00
|
|
|
// Crude data integrity check:
|
|
|
|
if (a_Header->m_UncompressedSize < cChunk::c_BlockDataSize)
|
|
|
|
{
|
|
|
|
LOGWARNING("Chunk [%d, %d] has too short decompressed data (%d out of %d needed), erasing",
|
|
|
|
a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ,
|
|
|
|
a_Header->m_UncompressedSize, cChunk::c_BlockDataSize
|
|
|
|
);
|
|
|
|
EraseChunk(a_Chunk);
|
|
|
|
m_NumDirty++;
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-02-13 16:47:03 -05:00
|
|
|
// Decompress the data:
|
2012-02-16 08:42:35 -05:00
|
|
|
AString UncompressedData;
|
|
|
|
int errorcode = UncompressString(m_DataContents.data() + a_Offset, a_Header->m_CompressedSize, UncompressedData, a_Header->m_UncompressedSize);
|
2012-02-13 16:47:03 -05:00
|
|
|
if (errorcode != Z_OK)
|
|
|
|
{
|
|
|
|
LOGERROR("Error %d decompressing data for chunk [%d, %d] from file \"%s\"",
|
|
|
|
errorcode,
|
2012-02-16 08:42:35 -05:00
|
|
|
a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ,
|
2012-02-13 16:47:03 -05:00
|
|
|
m_FileName.c_str()
|
|
|
|
);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
if (a_Header->m_UncompressedSize != (int)UncompressedData.size())
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
LOGWARNING("Uncompressed data size differs (exp %d, got %d) for chunk [%d, %d] from file \"%s\"",
|
2012-02-16 08:42:35 -05:00
|
|
|
a_Header->m_UncompressedSize, UncompressedData.size(),
|
|
|
|
a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ,
|
2012-02-13 16:47:03 -05:00
|
|
|
m_FileName.c_str()
|
|
|
|
);
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
cEntityList Entities;
|
|
|
|
cBlockEntityList BlockEntities;
|
|
|
|
|
|
|
|
if (a_Header->m_UncompressedSize > cChunk::c_BlockDataSize ) // We gots some extra data :D
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
LOGINFO("Parsing trailing JSON");
|
|
|
|
Json::Value root; // will contain the root value after parsing.
|
|
|
|
Json::Reader reader;
|
2012-02-16 08:42:35 -05:00
|
|
|
if ( !reader.parse( UncompressedData.data() + cChunk::c_BlockDataSize, root, false ) )
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
LOGERROR("Failed to parse trailing JSON!");
|
|
|
|
}
|
|
|
|
else
|
|
|
|
{
|
2012-02-16 08:42:35 -05:00
|
|
|
LoadEntitiesFromJson(root, Entities, BlockEntities, a_World);
|
2012-02-13 16:47:03 -05:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
a_World->ChunkDataLoaded(a_Chunk.m_ChunkX, 0, a_Chunk.m_ChunkZ, UncompressedData.data(), Entities, BlockEntities);
|
|
|
|
|
2012-02-13 16:47:03 -05:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
void cWSSCompact::cPAKFile::EraseChunk(const cChunkCoords & a_Chunk)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
2012-02-16 08:42:35 -05:00
|
|
|
int ChunkX = a_Chunk.m_ChunkX;
|
|
|
|
int ChunkZ = a_Chunk.m_ChunkZ;
|
2012-02-13 16:47:03 -05:00
|
|
|
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;
|
|
|
|
}
|
|
|
|
Offset += (*itr)->m_CompressedSize;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
bool cWSSCompact::cPAKFile::SaveChunkToData(const cChunkCoords & a_Chunk, cWorld * a_World)
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
// Serialize the chunk:
|
2012-02-16 08:42:35 -05:00
|
|
|
cJsonChunkSerializer Serializer;
|
|
|
|
a_World->GetChunkData(a_Chunk.m_ChunkX, 0, a_Chunk.m_ChunkZ, &Serializer);
|
|
|
|
if (Serializer.GetBlockData().empty())
|
|
|
|
{
|
|
|
|
// Chunk not valid
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-02-13 16:47:03 -05:00
|
|
|
AString Data;
|
2012-02-16 08:42:35 -05:00
|
|
|
std::swap(Serializer.GetBlockData(), Data);
|
|
|
|
if (Serializer.HasJsonData())
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
AString JsonData;
|
|
|
|
Json::StyledWriter writer;
|
2012-02-16 08:42:35 -05:00
|
|
|
JsonData = writer.write(Serializer.GetRoot());
|
2012-02-13 16:47:03 -05:00
|
|
|
Data.append(JsonData);
|
|
|
|
}
|
|
|
|
|
|
|
|
// Compress the data:
|
2012-02-16 08:42:35 -05:00
|
|
|
AString CompressedData;
|
|
|
|
int errorcode = CompressString(Data.data(), Data.size(), CompressedData);
|
2012-02-13 16:47:03 -05:00
|
|
|
if ( errorcode != Z_OK )
|
|
|
|
{
|
2012-02-16 08:42:35 -05:00
|
|
|
LOGERROR("Error %i compressing data for chunk [%d, %d]", errorcode, a_Chunk.m_ChunkX, a_Chunk.m_ChunkZ);
|
2012-02-13 16:47:03 -05:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
// Erase any existing data for the chunk:
|
|
|
|
EraseChunk(a_Chunk);
|
|
|
|
|
2012-02-13 16:47:03 -05:00
|
|
|
// Save the header:
|
|
|
|
sChunkHeader * Header = new sChunkHeader;
|
|
|
|
if (Header == NULL)
|
|
|
|
{
|
|
|
|
return false;
|
|
|
|
}
|
2012-02-16 08:42:35 -05:00
|
|
|
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();
|
2012-02-13 16:47:03 -05:00
|
|
|
m_ChunkHeaders.push_back(Header);
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
m_DataContents.append(CompressedData.data(), CompressedData.size());
|
2012-02-13 16:47:03 -05:00
|
|
|
|
|
|
|
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;
|
|
|
|
}
|
|
|
|
|
|
|
|
char PakVersion = 1;
|
|
|
|
WRITE(PakVersion);
|
|
|
|
char ChunkVersion = 1;
|
|
|
|
WRITE(ChunkVersion);
|
|
|
|
short NumChunks = (short)m_ChunkHeaders.size();
|
|
|
|
WRITE(NumChunks);
|
|
|
|
for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
|
|
|
|
{
|
|
|
|
WRITE(**itr);
|
|
|
|
}
|
2012-02-16 08:42:35 -05:00
|
|
|
if (f.Write(m_DataContents.data(), m_DataContents.size()) != (int)m_DataContents.size())
|
2012-02-13 16:47:03 -05:00
|
|
|
{
|
|
|
|
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;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-02-16 08:42:35 -05:00
|
|
|
|
|
|
|
void cWSSCompact::cPAKFile::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[]
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|