snes9x/jma/jma.cpp

/*
Copyright (C) 2005-2006 NSRT Team ( http://nsrt.edgeemu.com )

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
version 2 as published by the Free Software Foundation.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/

#include <sstream>
#include "jma.h"
using namespace std;

#include "portable.h"
#include "7z.h"
#include "crc32.h"

namespace JMA
{
  const char jma_magic[] = { 'J', 'M', 'A', 0, 'N' };
  const unsigned int jma_header_length = 5;
  const unsigned char jma_version = 1;
  const unsigned int jma_version_length = 1;
  const unsigned int jma_total_header_length = jma_header_length + jma_version_length + UINT_SIZE;

  //Convert DOS/zip/JMA integer time to to time_t
  time_t uint_to_time(unsigned short date, unsigned short time)
  {
    tm formatted_time;

    formatted_time.tm_mday = date & 0x1F;
    formatted_time.tm_mon  = ((date >> 5) & 0xF) - 1;
    formatted_time.tm_year = ((date >> 9) & 0x7f) + 80;
    formatted_time.tm_sec  = (time & 0x1F) * 2;
    formatted_time.tm_min  = (time >> 5) & 0x3F;
    formatted_time.tm_hour = (time >> 11) & 0x1F;

    return(mktime(&formatted_time));
  }


  //Retreive the file block, what else?
  void jma_open::retrieve_file_block()
  {
    unsigned char uint_buffer[UINT_SIZE];
    unsigned char ushort_buffer[USHORT_SIZE];

    //File block size is the last UINT in the file
    stream.seekg(-UINT_SIZE,ios::end);
    stream.read((char *)uint_buffer, UINT_SIZE);
    size_t file_block_size = charp_to_uint(uint_buffer);

    //Currently at the end of the file, so that's the file size
    size_t jma_file_size = (size_t) stream.tellg();

    //The file block can't be larger than the JMA file without it's header.
    //This if can probably be improved
    if (file_block_size >= jma_file_size-jma_total_header_length)
    {
      throw(JMA_BAD_FILE);
    }

    //Seek to before file block so we can read the file block
    stream.seekg(-((int)file_block_size+UINT_SIZE),ios::end);

    //This is needed if the file block is compressed
    stringstream decompressed_file_block;
    //Pointer to where to read file block from (file or decompressed buffer)
    istream *file_block_stream;

    //Setup file info buffer and byte to read with
    jma_file_info file_info;
    char byte;

    stream.get(byte);
    if (!byte) //If file block is compressed
    {
      //Compressed size isn't counting the byte we just read or the UINT for compressed size
      size_t compressed_size = file_block_size - (1+UINT_SIZE);

      //Read decompressed size / true file block size
      stream.read((char *)uint_buffer, UINT_SIZE);
      file_block_size = charp_to_uint(uint_buffer);

      //Setup access methods for decompression
      ISequentialInStream_Istream compressed_data(stream);
      ISequentialOutStream_Ostream decompressed_data(decompressed_file_block);

      //Decompress the data
      if (!decompress_lzma_7z(compressed_data, compressed_size, decompressed_data, file_block_size))
      {
        throw(JMA_DECOMPRESS_FAILED);
      }

      //Go to beginning, setup pointer to buffer
      decompressed_file_block.seekg(0, ios::beg);
      file_block_stream = &decompressed_file_block;
    }
    else
    {
      stream.putback(byte); //Putback byte, byte is part of filename, not compressed indicator
      file_block_stream = &stream;
    }


    //Minimum file name length is 2 bytes, a char and a null
    //Minimum comment length is 1 byte, a null
    //There are currently 2 UINTs and 2 USHORTs per file
    while (file_block_size >= 2+1+UINT_SIZE*2+USHORT_SIZE*2) //This does allow for a gap, but that's okay
    {
      //First stored in the file block is the file name null terminated
      file_info.name = "";

      file_block_stream->get(byte);
      while (byte)
      {
        file_info.name += byte;
        file_block_stream->get(byte);
      }

      //There must be a file name or the file is bad
      if (!file_info.name.length())
      {
        throw(JMA_BAD_FILE);
      }

      //Same trick as above for the comment
      file_info.comment = "";

      file_block_stream->get(byte);
      while (byte)
      {
        file_info.comment += byte;
        file_block_stream->get(byte);
      }

      //Next is a UINT representing the file's size
      file_block_stream->read((char *)uint_buffer, UINT_SIZE);
      file_info.size = charp_to_uint(uint_buffer);

      //Followed by CRC32
      file_block_stream->read((char *)uint_buffer, UINT_SIZE);
      file_info.crc32 = charp_to_uint(uint_buffer);

      //Special USHORT representation of file's date
      file_block_stream->read((char *)ushort_buffer, USHORT_SIZE);
      file_info.date = charp_to_ushort(ushort_buffer);

      //Special USHORT representation of file's time
      file_block_stream->read((char *)ushort_buffer, USHORT_SIZE);
      file_info.time = charp_to_ushort(ushort_buffer);

      file_info.buffer = 0; //Pointing to null till we decompress files

      files.push_back(file_info); //Put file info into our structure

      //Subtract size of the file info we just read
      file_block_size -= file_info.name.length()+file_info.comment.length()+2+UINT_SIZE*2+USHORT_SIZE*2;
    }
  }

  //Constructor for opening JMA files for reading
  jma_open::jma_open(const char *compressed_file_name)
  {
    decompressed_buffer = 0;
    compressed_buffer = 0;

    stream.open(compressed_file_name, ios::in | ios::binary);
    if (!stream.is_open())
    {
      throw(JMA_NO_OPEN);
    }

    //Header is "JMA\0N"
    unsigned char header[jma_header_length];
    stream.read((char *)header, jma_header_length);
    if (memcmp(jma_magic, header, jma_header_length))
    {
      throw(JMA_BAD_FILE);
    }

    //Not the cleanest code but logical
    stream.read((char *)header, 5);
    if (*header <= jma_version)
    {
      chunk_size = charp_to_uint(header+1); //Chunk size is a UINT that follows version #
      retrieve_file_block();
    }
    else
    {
      throw(JMA_UNSUPPORTED_VERSION);
    }
  }

  //Destructor only has to close the stream if neccesary
  jma_open::~jma_open()
  {
    if (stream.is_open())
    {
      stream.close();
    }
  }

  //Return a vector containing useful info about the files in the JMA
  vector<jma_public_file_info> jma_open::get_files_info()
  {
    vector<jma_public_file_info> file_info_vector;
    jma_public_file_info file_info;

    for (vector<jma_file_info>::iterator i = files.begin(); i != files.end(); i++)
    {
      file_info.name = i->name;
      file_info.comment = i->comment;
      file_info.size = i->size;
      file_info.datetime = uint_to_time(i->date, i->time);
      file_info.crc32 = i->crc32;
      file_info_vector.push_back(file_info);
    }

    return(file_info_vector);
  }

  //Skip forward a given number of chunks
  void jma_open::chunk_seek(unsigned int chunk_num)
  {
    //Check the stream is open
    if (!stream.is_open())
    {
      throw(JMA_NO_OPEN);
    }

    //Clear possible errors so the seek will work
    stream.clear();

    //Move forward over header
    stream.seekg(jma_total_header_length, ios::beg);

    unsigned char int4_buffer[UINT_SIZE];

    while (chunk_num--)
    {
      //Read in size of chunk
      stream.read((char *)int4_buffer, UINT_SIZE);

      //Skip chunk plus it's CRC32
      stream.seekg(charp_to_uint(int4_buffer)+UINT_SIZE, ios::cur);
    }
  }

  //Return a vector of pointers to each file in the JMA, the buffer to hold all the files
  //must be initilized outside.
  vector<unsigned char *> jma_open::get_all_files(unsigned char *buffer)
  {
    //If there's no stream we can't read from it, so exit
    if (!stream.is_open())
    {
      throw(JMA_NO_OPEN);
    }

    //Seek to the first chunk
    chunk_seek(0);

    //Set the buffer that decompressed data goes to
    decompressed_buffer = buffer;

    //If the JMA is not solid
    if (chunk_size)
    {
      unsigned char int4_buffer[UINT_SIZE];
      size_t size = get_total_size(files);

      //For each chunk in the file...
      for (size_t remaining_size = size; remaining_size; remaining_size -= chunk_size)
      {
        //Read the compressed size
        stream.read((char *)int4_buffer, UINT_SIZE);
        size_t compressed_size = charp_to_uint(int4_buffer);

        //Allocate memory of the correct size to hold the compressed data in the JMA
        //Throw error on failure as that is unrecoverable from
        try
        {
          compressed_buffer = new unsigned char[compressed_size];
        }
        catch (bad_alloc xa)
        {
          throw(JMA_NO_MEM_ALLOC);
        }

        //Read all the compressed data in
        stream.read((char *)compressed_buffer, compressed_size);

        //Read the expected CRC of compressed data from the file
        stream.read((char *)int4_buffer, UINT_SIZE);

        //If it doesn't match, throw error and cleanup memory
        if (CRC32lib::CRC32(compressed_buffer, compressed_size) != charp_to_uint(int4_buffer))
        {
          delete[] compressed_buffer;
          throw(JMA_BAD_FILE);
        }

        //Decompress the data, cleanup memory on failure
        if (!decompress_lzma_7z(compressed_buffer, compressed_size,
                                decompressed_buffer+size-remaining_size,
                                (remaining_size > chunk_size) ? chunk_size : remaining_size))
        {
          delete[] compressed_buffer;
          throw(JMA_DECOMPRESS_FAILED);
        }
        delete[] compressed_buffer;

        if (remaining_size <= chunk_size) //If we just decompressed the remainder
        {
          break;
        }
      }
    }
    else //Solidly compressed JMA
    {
      unsigned char int4_buffer[UINT_SIZE];

      //Read the size of the compressed data
      stream.read((char *)int4_buffer, UINT_SIZE);
      size_t compressed_size = charp_to_uint(int4_buffer);

      //Get decompressed size
      size_t size = get_total_size(files);

      //Setup access methods for decompression
      ISequentialInStream_Istream compressed_data(stream);
      ISequentialOutStream_Array decompressed_data(reinterpret_cast<char*>(decompressed_buffer), size);

      //Decompress the data
      if (!decompress_lzma_7z(compressed_data, compressed_size, decompressed_data, size))
      {
        throw(JMA_DECOMPRESS_FAILED);
      }

      /*
      //Allocate memory of the right size to hold the compressed data in the JMA
      try
      {
        compressed_buffer = new unsigned char[compressed_size];
      }
      catch (bad_alloc xa)
      {
        throw(JMA_NO_MEM_ALLOC);
      }

      //Copy the compressed data into memory
      stream.read((char *)compressed_buffer, compressed_size);
      size_t size = get_total_size(files);

      //Read the CRC of the compressed data
      stream.read((char *)int4_buffer, UINT_SIZE);

      //If it doesn't match, complain
      if (CRC32lib::CRC32(compressed_buffer, compressed_size) != charp_to_uint(int4_buffer))
      {
        delete[] compressed_buffer;
        throw(JMA_BAD_FILE);
      }

      //Decompress the data
      if (!decompress_lzma_7z(compressed_buffer, compressed_size, decompressed_buffer, size))
      {
        delete[] compressed_buffer;
        throw(JMA_DECOMPRESS_FAILED);
      }
      delete[] compressed_buffer;
      */
    }

    vector<unsigned char *> file_pointers;
    size_t size = 0;

    //For each file, add it's pointer to the vector, size is pointer offset in the buffer
    for (vector<jma_file_info>::iterator i = files.begin(); i != files.end(); i++)
    {
      i->buffer = decompressed_buffer+size;
      file_pointers.push_back(decompressed_buffer+size);
      size += i->size;
    }

    //Return the vector of pointers
    return(file_pointers);
  }

  //Extracts the file with a given name found in the archive to the given buffer
  void jma_open::extract_file(string& name, unsigned char *buffer)
  {
    if (!stream.is_open())
    {
      throw(JMA_NO_OPEN);
    }

    size_t size_to_skip = 0;
    size_t our_file_size = 0;

    //Search through the vector of file information
    for (vector<jma_file_info>::iterator i = files.begin(); i != files.end(); i++)
    {
      if (i->name == name)
      {
        //Set the variable so we can tell we found it
        our_file_size = i->size;
        break;
      }

      //Keep a running total of size
      size_to_skip += i->size;
    }

    if (!our_file_size) //File with the specified name was not found in the archive
    {
      throw(JMA_FILE_NOT_FOUND);
    }

    //If the JMA only contains one file, we can skip a lot of overhead
    if (files.size() == 1)
    {
      get_all_files(buffer);
      return;
    }

    if (chunk_size) //we are using non-solid archive..
    {
      unsigned int chunks_to_skip = size_to_skip / chunk_size;

      //skip over requisite number of chunks
      chunk_seek(chunks_to_skip);

      //Allocate memory for compressed and decompressed data
      unsigned char *comp_buffer = 0, *decomp_buffer = 0;
      try
      {
        //Compressed data size is <= non compressed size
        unsigned char *combined_buffer = new unsigned char[chunk_size*2];
        comp_buffer = combined_buffer;
        decomp_buffer = combined_buffer+chunk_size;
      }
      catch (bad_alloc xa)
      {
        throw(JMA_NO_MEM_ALLOC);
      }

      size_t first_chunk_offset = size_to_skip % chunk_size;
      unsigned char int4_buffer[UINT_SIZE];
      for (size_t i = 0; i < our_file_size;)
      {
        //Get size
        stream.read((char *)int4_buffer, UINT_SIZE);
        size_t compressed_size = charp_to_uint(int4_buffer);

        //Read all the compressed data in
        stream.read((char *)comp_buffer, compressed_size);

        //Read the CRC of the compressed data
        stream.read((char *)int4_buffer, UINT_SIZE);

        //If it doesn't match, complain
        if (CRC32lib::CRC32(comp_buffer, compressed_size) != charp_to_uint(int4_buffer))
        {
          delete[] comp_buffer;
          throw(JMA_BAD_FILE);
        }

        //Decompress chunk
        if (!decompress_lzma_7z(comp_buffer, compressed_size, decomp_buffer, chunk_size))
        {
          delete[] comp_buffer;
          throw(JMA_DECOMPRESS_FAILED);
        }

        size_t copy_amount = our_file_size-i > chunk_size-first_chunk_offset ? chunk_size-first_chunk_offset : our_file_size-i;

        memcpy(buffer+i, decomp_buffer+first_chunk_offset, copy_amount);
        first_chunk_offset = 0; //Set to zero since this is only for the first iteration
        i += copy_amount;
      }
      delete[] comp_buffer;
    }
    else //Solid JMA
    {
      unsigned char *decomp_buffer = 0;
      try
      {
        decomp_buffer = new unsigned char[get_total_size(files)];
      }
      catch (bad_alloc xa)
      {
        throw(JMA_NO_MEM_ALLOC);
      }

      get_all_files(decomp_buffer);

      memcpy(buffer, decomp_buffer+size_to_skip, our_file_size);

      delete[] decomp_buffer;
    }
  }

  bool jma_open::is_solid()
  {
    return(chunk_size ? false : true);
  }

  const char *jma_error_text(jma_errors error)
  {
    switch (error)
    {
      case JMA_NO_CREATE:
        return("JMA could not be created");

      case JMA_NO_MEM_ALLOC:
        return("Memory for JMA could be allocated");

      case JMA_NO_OPEN:
        return("JMA could not be opened");

      case JMA_BAD_FILE:
        return("Invalid/Corrupt JMA");

      case JMA_UNSUPPORTED_VERSION:
        return("JMA version not supported");

      case JMA_COMPRESS_FAILED:
        return("JMA compression failed");

      case JMA_DECOMPRESS_FAILED:
        return("JMA decompression failed");

      case JMA_FILE_NOT_FOUND:
        return("File not found in JMA");
    }
    return("Unknown error");
  }

}