summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
-rw-r--r--source/WSSCompact.cpp170
-rw-r--r--source/WSSCompact.h9
-rw-r--r--source/cChunk.h2
3 files changed, 165 insertions, 16 deletions
diff --git a/source/WSSCompact.cpp b/source/WSSCompact.cpp
index bb47ac0cc..195a450ec 100644
--- a/source/WSSCompact.cpp
+++ b/source/WSSCompact.cpp
@@ -269,7 +269,9 @@ cWSSCompact::cPAKFile::cPAKFile(const AString & a_FileName, int a_LayerX, int a_
m_FileName(a_FileName),
m_LayerX(a_LayerX),
m_LayerZ(a_LayerZ),
- m_NumDirty(0)
+ m_NumDirty(0),
+ m_ChunkVersion( PAK_VERSION ), // Init with latest version
+ m_PakVersion( CHUNK_VERSION )
{
cFile f;
if (!f.Open(m_FileName, cFile::fmRead))
@@ -278,21 +280,26 @@ cWSSCompact::cPAKFile::cPAKFile(const AString & a_FileName, int a_LayerX, int a_
}
// Read headers:
- char PakVersion = 0;
- READ(PakVersion);
- if (PakVersion != 1)
+ READ(m_PakVersion);
+ if (m_PakVersion != 1)
{
- LOGERROR("File \"%s\" is in an unknown pak format (%d)", m_FileName.c_str(), PakVersion);
+ LOGERROR("File \"%s\" is in an unknown pak format (%d)", m_FileName.c_str(), m_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);
+ READ(m_ChunkVersion);
+ switch( m_ChunkVersion )
+ {
+ case 1:
+ m_ChunkSize.Set(16, 128, 16);
+ break;
+ case 2:
+ 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);
@@ -311,6 +318,11 @@ cWSSCompact::cPAKFile::cPAKFile(const AString & a_FileName, int a_LayerX, int a_
LOGERROR("Cannot read file \"%s\" contents", m_FileName.c_str());
return;
}
+
+ if( m_ChunkVersion == 1 ) // Convert chunks to version 2
+ {
+ UpdateChunk1To2();
+ }
}
@@ -380,6 +392,136 @@ bool cWSSCompact::cPAKFile::SaveChunk(const cChunkCoords & a_Chunk, cWorld * a_W
+void cWSSCompact::cPAKFile::UpdateChunk1To2()
+{
+ LOGINFO("Updating \"%s\" version 1 to version 2", m_FileName.c_str() );
+ int Offset = 0;
+ AString NewDataContents;
+ for (sChunkHeaders::iterator itr = m_ChunkHeaders.begin(); itr != m_ChunkHeaders.end(); ++itr)
+ {
+ sChunkHeader * Header = *itr;
+
+ LOGINFO("Updating \"%s\" version 1 to version 2: Updating chunk [%d, %d]", m_FileName.c_str(), Header->m_ChunkX, Header->m_ChunkZ );
+
+ 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
+ AString ConvertedData;
+ 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.push_back( UncompressedData[y + z*128 + x*128*16 + InChunkOffset] );
+ }
+ // Add 128 empty blocks after an old y column
+ ConvertedData.append( 128, E_BLOCK_AIR );
+ }
+ 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.push_back( UncompressedData[y + z*64 + x*64*16 + InChunkOffset] );
+ }
+ ConvertedData.append( 64, 0 );
+ }
+ 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.push_back( UncompressedData[y + z*64 + x*64*16 + InChunkOffset] );
+ }
+ ConvertedData.append( 64, 0 );
+ }
+ 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.push_back( UncompressedData[y + z*64 + x*64*16 + InChunkOffset] );
+ }
+ ConvertedData.append( 64, 0 );
+ }
+ InChunkOffset += (16*128*16)/2;
+ // Add JSON data afterwards
+ if( UncompressedData.size() > InChunkOffset )
+ {
+ ConvertedData.append( UncompressedData.begin() + InChunkOffset, UncompressedData.end() );
+ }
+
+ // Re-compress data
+ AString CompressedData;
+ {
+ int errorcode = CompressString(ConvertedData.data(), ConvertedData.size(), CompressedData);
+ if (errorcode != Z_OK)
+ {
+ LOGERROR("Error %d compressing data for chunk [%d, %d]",
+ errorcode,
+ Header->m_ChunkX, Header->m_ChunkZ
+ );
+ continue;
+ }
+ }
+
+ Header->m_UncompressedSize = ConvertedData.size();
+ Header->m_CompressedSize = CompressedData.size();
+
+
+ NewDataContents.append( CompressedData );
+ }
+
+ // Done converting
+ m_DataContents = NewDataContents;
+ m_ChunkVersion = 2;
+}
+
+
+
+
+
bool cWSSCompact::LoadChunkFromData(const cChunkCoords & a_Chunk, int & a_UncompressedSize, const AString & a_Data, cWorld * a_World)
{
// Crude data integrity check:
@@ -539,10 +681,8 @@ void cWSSCompact::cPAKFile::SynchronizeFile(void)
return;
}
- char PakVersion = 1;
- WRITE(PakVersion);
- char ChunkVersion = 1;
- WRITE(ChunkVersion);
+ 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)
diff --git a/source/WSSCompact.h b/source/WSSCompact.h
index 1cb8d161b..c42c8c5b9 100644
--- a/source/WSSCompact.h
+++ b/source/WSSCompact.h
@@ -12,6 +12,7 @@
#define WSSCOMPACT_H_INCLUDED
#include "WorldStorage.h"
+#include "Vector3i.h"
@@ -47,6 +48,8 @@ protected:
int GetLayerX(void) const {return m_LayerX; }
int GetLayerZ(void) const {return m_LayerZ; }
+ static const int PAK_VERSION = 1;
+ static const int CHUNK_VERSION = 2;
protected:
AString m_FileName;
@@ -57,10 +60,16 @@ protected:
AString m_DataContents; // Data contents of the file, cached
int m_NumDirty; // Number of chunks that were written into m_DataContents but not into the file
+
+ Vector3i m_ChunkSize; // Is related to m_ChunkVersion
+ char m_ChunkVersion;
+ char m_PakVersion;
bool LoadChunk(const cChunkCoords & a_Chunk, int a_Offset, sChunkHeader * a_Header, cWorld * a_World);
bool SaveChunkToData(const cChunkCoords & a_Chunk, cWorld * a_World); // Saves the chunk to m_DataContents, updates headers and m_NumDirty
void SynchronizeFile(void); // Writes m_DataContents along with the headers to file, resets m_NumDirty
+
+ void UpdateChunk1To2(void);
} ;
typedef std::list<cPAKFile *> cPAKFiles;
diff --git a/source/cChunk.h b/source/cChunk.h
index f2e623f6e..dfbbd0c98 100644
--- a/source/cChunk.h
+++ b/source/cChunk.h
@@ -113,7 +113,7 @@ class cChunk
{
public:
static const int c_ChunkWidth = 16;
- static const int c_ChunkHeight = 128; //256;
+ static const int c_ChunkHeight = 256;
static const int c_NumBlocks = c_ChunkWidth * c_ChunkHeight * c_ChunkWidth;
static const int c_BlockDataSize = c_NumBlocks * 2 + (c_NumBlocks/2); // 2.5 * numblocks