/***************************************************************************

    CHD compression frontend

****************************************************************************

    Copyright Aaron Giles
    All rights reserved.

    Redistribution and use in source and binary forms, with or without
    modification, are permitted provided that the following conditions are
    met:

        * Redistributions of source code must retain the above copyright
          notice, this list of conditions and the following disclaimer.
        * Redistributions in binary form must reproduce the above copyright
          notice, this list of conditions and the following disclaimer in
          the documentation and/or other materials provided with the
          distribution.
        * Neither the name 'MAME' nor the names of its contributors may be
          used to endorse or promote products derived from this software
          without specific prior written permission.

    THIS SOFTWARE IS PROVIDED BY AARON GILES ''AS IS'' AND ANY EXPRESS OR
    IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
    WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
    DISCLAIMED. IN NO EVENT SHALL AARON GILES BE LIABLE FOR ANY DIRECT,
    INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
    (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
    SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
    HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
    STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
    IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
    POSSIBILITY OF SUCH DAMAGE.

****************************************************************************/

#include "osdcore.h"
#include "corefile.h"
#include "chdcd.h"
#include "aviio.h"
#include "avhuff.h"
#include "bitmap.h"
#include "md5.h"
#include "sha1.h"
#include "vbiparse.h"
#include "tagmap.h"
#include <stdarg.h>
#include <stdlib.h>
#include <stdio.h>
#include <time.h>
#include <ctype.h>
#include <new>



//**************************************************************************
//  CONSTANTS & DEFINES
//**************************************************************************

// default hard disk sector size
const UINT32 IDE_SECTOR_SIZE = 512;

// temporary input buffer size
const UINT32 TEMP_BUFFER_SIZE = 32 * 1024 * 1024;

// modes
const int MODE_NORMAL = 0;
const int MODE_CUEBIN = 1;
const int MODE_GDI = 2;

// command modifier
#define REQUIRED "~"

// command strings
#define COMMAND_HELP "help"
#define COMMAND_INFO "info"
#define COMMAND_VERIFY "verify"
#define COMMAND_CREATE_RAW "createraw"
#define COMMAND_CREATE_HD "createhd"
#define COMMAND_CREATE_CD "createcd"
#define COMMAND_CREATE_LD "createld"
#define COMMAND_EXTRACT_RAW "extractraw"
#define COMMAND_EXTRACT_HD "extracthd"
#define COMMAND_EXTRACT_CD "extractcd"
#define COMMAND_EXTRACT_LD "extractld"
#define COMMAND_COPY "copy"
#define COMMAND_ADD_METADATA "addmeta"
#define COMMAND_DEL_METADATA "delmeta"
#define COMMAND_DUMP_METADATA "dumpmeta"

// option strings
#define OPTION_INPUT "input"
#define OPTION_OUTPUT "output"
#define OPTION_OUTPUT_BIN "outputbin"
#define OPTION_OUTPUT_FORCE "force"
#define OPTION_INPUT_START_BYTE "inputstartbyte"
#define OPTION_INPUT_START_HUNK "inputstarthunk"
#define OPTION_INPUT_START_FRAME "inputstartframe"
#define OPTION_INPUT_LENGTH_BYTES "inputbytes"
#define OPTION_INPUT_LENGTH_HUNKS "inputhunks"
#define OPTION_INPUT_LENGTH_FRAMES "inputframes"
#define OPTION_HUNK_SIZE "hunksize"
#define OPTION_UNIT_SIZE "unitsize"
#define OPTION_COMPRESSION "compression"
#define OPTION_INPUT_PARENT "inputparent"
#define OPTION_OUTPUT_PARENT "outputparent"
#define OPTION_IDENT "ident"
#define OPTION_CHS "chs"
#define OPTION_SECTOR_SIZE "sectorsize"
#define OPTION_TAG "tag"
#define OPTION_INDEX "index"
#define OPTION_VALUE_TEXT "valuetext"
#define OPTION_VALUE_FILE "valuefile"
#define OPTION_NO_CHECKSUM "nochecksum"
#define OPTION_VERBOSE "verbose"
#define OPTION_FIX "fix"
#define OPTION_NUMPROCESSORS "numprocessors"


//**************************************************************************
//  FUNCTION PROTOTYPES
//**************************************************************************

typedef tagmap_t<astring *> parameters_t;

static void report_error(int error, const char *format, ...);
static void do_info(parameters_t &params);
static void do_verify(parameters_t &params);
static void do_create_raw(parameters_t &params);
static void do_create_hd(parameters_t &params);
static void do_create_cd(parameters_t &params);
static void do_create_ld(parameters_t &params);
static void do_copy(parameters_t &params);
static void do_extract_raw(parameters_t &params);
static void do_extract_cd(parameters_t &params);
static void do_extract_ld(parameters_t &params);
static void do_add_metadata(parameters_t &params);
static void do_del_metadata(parameters_t &params);
static void do_dump_metadata(parameters_t &params);



//**************************************************************************
//  TYPE DEFINITIONS
//**************************************************************************

// ======================> option_description

struct option_description
{
	const char *name;
	const char *shortname;
	bool parameter;
	const char *description;
};


// ======================> command_description

struct command_description
{
	const char *name;
	void (*handler)(parameters_t &);
	const char *description;
	const char *valid_options[16];
};


// ======================> avi_info

struct avi_info
{
	UINT32 fps_times_1million;
	UINT32 width;
	UINT32 height;
	bool interlaced;
	UINT32 channels;
	UINT32 rate;
	UINT32 max_samples_per_frame;
	UINT32 bytes_per_frame;
};


// ======================> metadata_index_info

struct metadata_index_info
{
	chd_metadata_tag	tag;
	UINT32				index;
};


// ======================> fatal_error

class fatal_error : public std::exception
{
public:
	fatal_error(int error)
		: m_error(error) { }

	int error() const { return m_error; }

private:
	int m_error;
};


// ======================> chd_rawfile_compressor

class chd_rawfile_compressor : public chd_file_compressor
{
public:
	// construction/destruction
	chd_rawfile_compressor(core_file *file, UINT64 offset = 0, UINT64 maxoffset = ~0)
		: m_file(file),
		  m_offset(offset),
		  m_maxoffset(MIN(maxoffset, (file != NULL) ? core_fsize(file) : 0)) { }

	// read interface
	virtual UINT32 read_data(void *dest, UINT64 offset, UINT32 length)
	{
		offset += m_offset;
		if (offset >= m_maxoffset)
			return 0;
		if (offset + length > m_maxoffset)
			length = m_maxoffset - offset;
		core_fseek(m_file, offset, SEEK_SET);
		return core_fread(m_file, dest, length);
	}

private:
	// internal state
	core_file *		m_file;
	UINT64			m_offset;
	UINT64			m_maxoffset;
};


// ======================> chd_chdfile_compressor

class chd_chdfile_compressor : public chd_file_compressor
{
public:
	// construction/destruction
	chd_chdfile_compressor(chd_file &file, UINT64 offset = 0, UINT64 maxoffset = ~0)
		: m_file(file),
		  m_offset(offset),
		  m_maxoffset(MIN(maxoffset, file.logical_bytes())) { }

	// read interface
	virtual UINT32 read_data(void *dest, UINT64 offset, UINT32 length)
	{
		offset += m_offset;
		if (offset >= m_maxoffset)
			return 0;
		if (offset + length > m_maxoffset)
			length = m_maxoffset - offset;
		chd_error err = m_file.read_bytes(offset, dest, length);
		if (err != CHDERR_NONE)
			throw err;
		return length;
	}

private:
	// internal state
	chd_file &		m_file;
	UINT64			m_offset;
	UINT64			m_maxoffset;
};


// ======================> chd_cd_compressor

class chd_cd_compressor : public chd_file_compressor
{
public:
	// construction/destruction
	chd_cd_compressor(cdrom_toc &toc, chdcd_track_input_info &info)
		: m_file(NULL),
		  m_toc(toc),
		  m_info(info) { }

	~chd_cd_compressor()
	{
		if (m_file != NULL)
			core_fclose(m_file);
	}

	// read interface
	virtual UINT32 read_data(void *_dest, UINT64 offset, UINT32 length)
	{
		// verify assumptions made below
		assert(offset % CD_FRAME_SIZE == 0);
		assert(length % CD_FRAME_SIZE == 0);

		// initialize destination to 0 so that unused areas are filled
		UINT8 *dest = reinterpret_cast<UINT8 *>(_dest);
		memset(dest, 0, length);

		// find out which track we're starting in
		UINT64 startoffs = 0;
		UINT32 length_remaining = length;
		for (int tracknum = 0; tracknum < m_toc.numtrks; tracknum++)
		{
			const cdrom_track_info &trackinfo = m_toc.tracks[tracknum];
			UINT64 endoffs = startoffs + (trackinfo.frames + trackinfo.extraframes) * CD_FRAME_SIZE;
			if (offset >= startoffs && offset < endoffs)
			{
				// if we don't already have this file open, open it now
				if (m_file == NULL || m_lastfile != m_info.track[tracknum].fname)
				{
					if (m_file != NULL)
						core_fclose(m_file);
					m_lastfile = m_info.track[tracknum].fname;
					file_error filerr = core_fopen(m_lastfile, OPEN_FLAG_READ, &m_file);
					if (filerr != FILERR_NONE)
						report_error(1, "Error opening input file (%s)'", m_lastfile.cstr());
				}

				// iterate over frames
				UINT32 bytesperframe = trackinfo.datasize + trackinfo.subsize;
				UINT64 src_track_start = m_info.track[tracknum].offset;
				UINT64 src_track_end = src_track_start + bytesperframe * trackinfo.frames;
                UINT64 pad_track_start = src_track_end - (m_toc.tracks[tracknum].padframes * bytesperframe);
                while (length_remaining != 0 && offset < endoffs)
				{
					// determine start of current frame
					UINT64 src_frame_start = src_track_start + ((offset - startoffs) / CD_FRAME_SIZE) * bytesperframe;
					if (src_frame_start < src_track_end)
					{
						// read it in, or pad if we're into the padframes
                        if (src_frame_start >= pad_track_start)
                        {
                            memset(dest, 0, bytesperframe);
                        }
                        else
                        {
                            core_fseek(m_file, src_frame_start, SEEK_SET);
                            UINT32 count = core_fread(m_file, dest, bytesperframe);
                            if (count != bytesperframe)
                                report_error(1, "Error reading input file (%s)'", m_lastfile.cstr());
                        }

						// swap if appropriate
						if (m_info.track[tracknum].swap)
							for (UINT32 swapindex = 0; swapindex < 2352; swapindex += 2)
							{
								UINT8 temp = dest[swapindex];
								dest[swapindex] = dest[swapindex + 1];
								dest[swapindex + 1] = temp;
							}
					}

					// advance
					offset += CD_FRAME_SIZE;
					dest += CD_FRAME_SIZE;
					length_remaining -= CD_FRAME_SIZE;
					if (length_remaining == 0)
						break;
				}
			}

			// next track starts after the previous one
			startoffs = endoffs;
		}
		return length - length_remaining;
	}

private:
	// internal state
	astring						m_lastfile;
	core_file *					m_file;
	cdrom_toc &					m_toc;
	chdcd_track_input_info &	m_info;
};


// ======================> chd_avi_compressor

class chd_avi_compressor : public chd_file_compressor
{
public:
	// construction/destruction
	chd_avi_compressor(avi_file &file, avi_info &info, UINT32 first_frame, UINT32 num_frames)
		: m_file(file),
		  m_info(info),
		  m_bitmap(info.width, info.height * (info.interlaced ? 2 : 1)),
		  m_start_frame(first_frame),
		  m_frame_count(num_frames),
		  m_ldframedata(num_frames * VBI_PACKED_BYTES),
		  m_rawdata(info.bytes_per_frame) { }

	// getters
	const dynamic_buffer &ldframedata() const { return m_ldframedata; }

	// read interface
	virtual UINT32 read_data(void *_dest, UINT64 offset, UINT32 length)
	{
		UINT8 *dest = reinterpret_cast<UINT8 *>(_dest);
		int interlace_factor = m_info.interlaced ? 2 : 1;
		UINT32 length_remaining = length;

		// iterate over frames
		INT32 start_frame = offset / m_info.bytes_per_frame;
		INT32 end_frame = (offset + length - 1) / m_info.bytes_per_frame;
		for (INT32 framenum = start_frame; framenum <= end_frame; framenum++)
			if (framenum < m_frame_count)
			{
				// determine effective frame number and first/last samples
				INT32 effframe = m_start_frame + framenum;
				UINT32 first_sample = (UINT64(m_info.rate) * UINT64(effframe) * UINT64(1000000) + m_info.fps_times_1million - 1) / UINT64(m_info.fps_times_1million);
				UINT32 samples = (UINT64(m_info.rate) * UINT64(effframe + 1) * UINT64(1000000) + m_info.fps_times_1million - 1) / UINT64(m_info.fps_times_1million) - first_sample;

				// loop over channels and read the samples
				int channels = MIN(m_info.channels, ARRAY_LENGTH(m_audio));
				INT16 *samplesptr[ARRAY_LENGTH(m_audio)];
				for (int chnum = 0; chnum < channels; chnum++)
				{
					// read the sound samples
					m_audio[chnum].resize(samples);
					samplesptr[chnum] = m_audio[chnum];
					avi_error avierr = avi_read_sound_samples(&m_file, chnum, first_sample, samples, m_audio[chnum]);
					if (avierr != AVIERR_NONE)
						report_error(1, "Error reading audio samples %d-%d from channel %d: %s", first_sample, samples, chnum, avi_error_string(avierr));
				}

				// read the video data
				avi_error avierr = avi_read_video_frame(&m_file, effframe / interlace_factor, m_bitmap);
				if (avierr != AVIERR_NONE)
					report_error(1, "Error reading AVI frame %d: %s", effframe / interlace_factor, avi_error_string(avierr));
				bitmap_yuy16 subbitmap(&m_bitmap.pix(effframe % interlace_factor), m_bitmap.width(), m_bitmap.height() / interlace_factor, m_bitmap.rowpixels() * interlace_factor);

				// update metadata for this frame
				if (m_info.height == 524/2 || m_info.height == 624/2)
				{
					vbi_metadata vbi;
					vbi_parse_all(&subbitmap.pix16(0), subbitmap.rowpixels(), subbitmap.width(), 8, &vbi);
					vbi_metadata_pack(&m_ldframedata[framenum * VBI_PACKED_BYTES], framenum, &vbi);
				}

				// assemble the data into final form
				avhuff_error averr = avhuff_encoder::assemble_data(m_rawdata, subbitmap, channels, samples, samplesptr);
				if (averr != AVHERR_NONE)
					report_error(1, "Error assembling data for frame %d", framenum);
				if (m_rawdata.count() < m_info.bytes_per_frame)
				{
					UINT32 delta = m_info.bytes_per_frame - m_rawdata.count();
					m_rawdata.resize(m_info.bytes_per_frame, true);
					memset(&m_rawdata[m_info.bytes_per_frame - delta], 0, delta);
				}

				// copy to the destination
				UINT64 start_offset = UINT64(framenum) * UINT64(m_info.bytes_per_frame);
				UINT64 end_offset = start_offset + m_info.bytes_per_frame;
				UINT32 bytes_to_copy = MIN(length_remaining, end_offset - offset);
				memcpy(dest, &m_rawdata[offset - start_offset], bytes_to_copy);

				// advance
				offset += bytes_to_copy;
				dest += bytes_to_copy;
				length_remaining -= bytes_to_copy;
			}

		return length;
	}

private:
	// internal state
	avi_file &					m_file;
	avi_info &					m_info;
	bitmap_yuy16				m_bitmap;
	UINT32						m_start_frame;
	UINT32						m_frame_count;
	dynamic_array<INT16>		m_audio[8];
	dynamic_buffer				m_ldframedata;
	dynamic_buffer				m_rawdata;
};



//**************************************************************************
//  GLOBAL VARIABLES
//**************************************************************************

// timing
static clock_t lastprogress = 0;


// default compressors
static const chd_codec_type s_default_raw_compression[4] = { CHD_CODEC_LZMA, CHD_CODEC_ZLIB, CHD_CODEC_HUFFMAN, CHD_CODEC_FLAC };
static const chd_codec_type s_default_hd_compression[4] = { CHD_CODEC_LZMA, CHD_CODEC_ZLIB, CHD_CODEC_HUFFMAN, CHD_CODEC_FLAC };
static const chd_codec_type s_default_cd_compression[4] = { CHD_CODEC_CD_LZMA, CHD_CODEC_CD_ZLIB, CHD_CODEC_CD_FLAC };
static const chd_codec_type s_default_ld_compression[4] = { CHD_CODEC_AVHUFF };


// descriptions for each option
static const option_description s_options[] =
{
	{ OPTION_INPUT, 				"i",	true, " <filename>: input file name" },
	{ OPTION_INPUT_PARENT,			"ip",	true, " <filename>: parent file name for input CHD" },
	{ OPTION_OUTPUT,				"o",	true, " <filename>: output file name" },
	{ OPTION_OUTPUT_BIN,			"ob",	true, " <filename>: output file name for binary data" },
	{ OPTION_OUTPUT_FORCE,			"f",	false, ": force overwriting an existing file" },
	{ OPTION_OUTPUT_PARENT,			"op",	true, " <filename>: parent file name for output CHD" },
	{ OPTION_INPUT_START_BYTE,		"isb",	true, " <offset>: starting byte offset within the input" },
	{ OPTION_INPUT_START_HUNK,		"ish",	true, " <offset>: starting hunk offset within the input" },
	{ OPTION_INPUT_START_FRAME,		"isf",	true, " <offset>: starting frame within the input" },
	{ OPTION_INPUT_LENGTH_BYTES,	"ib",	true, " <length>: effective length of input in bytes" },
	{ OPTION_INPUT_LENGTH_HUNKS,	"ih",	true, " <length>: effective length of input in hunks" },
	{ OPTION_INPUT_LENGTH_FRAMES,	"if",	true, " <length>: effective length of input in frames" },
	{ OPTION_HUNK_SIZE,				"hs",	true, " <bytes>: size of each hunk, in bytes" },
	{ OPTION_UNIT_SIZE,				"us",	true, " <bytes>: size of each unit, in bytes" },
	{ OPTION_COMPRESSION,			"c",	true, " <none|type1[,type2[,...]]>: which compression codecs to use (up to 4)" },
	{ OPTION_IDENT,					"id",	true, " <filename>: name of ident file to provide CHS information" },
	{ OPTION_CHS,					"chs",	true, " <cylinders,heads,sectors>: specifies CHS values directly" },
	{ OPTION_SECTOR_SIZE,			"ss",	true, " <bytes>: size of each hard disk sector" },
	{ OPTION_TAG,					"t",	true, " <tag>: 4-character tag for metadata" },
	{ OPTION_INDEX,					"ix",	true, " <index>: indexed instance of this metadata tag" },
	{ OPTION_VALUE_TEXT,			"vt",	true, " <text>: text for the metadata" },
	{ OPTION_VALUE_FILE,			"vf",	true, " <file>: file containing data to add" },
	{ OPTION_NUMPROCESSORS,			"np",	true, " <processors>: limit the number of processors to use during compression" },
	{ OPTION_NO_CHECKSUM,			"nocs",	false, ": do not include this metadata information in the overall SHA-1" },
	{ OPTION_FIX,					"f",	false, ": fix the SHA-1 if it is incorrect" },
	{ OPTION_VERBOSE,				"v",	false, ": output additional information" },
};


// descriptions for each command
static const command_description s_commands[] =
{
	{ COMMAND_INFO, do_info, ": displays information about a CHD",
		{
			REQUIRED OPTION_INPUT,
			OPTION_VERBOSE
		}
	},

	{ COMMAND_VERIFY, do_verify, ": verifies a CHD's integrity",
		{
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_PARENT
		}
	},

	{ COMMAND_CREATE_RAW, do_create_raw, ": create a raw CHD from the input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_PARENT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_START_BYTE,
			OPTION_INPUT_START_HUNK,
			OPTION_INPUT_LENGTH_BYTES,
			OPTION_INPUT_LENGTH_HUNKS,
			REQUIRED OPTION_HUNK_SIZE,
			REQUIRED OPTION_UNIT_SIZE,
			OPTION_COMPRESSION,
			OPTION_NUMPROCESSORS
		}
	},

	{ COMMAND_CREATE_HD, do_create_hd, ": create a hard disk CHD from the input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_PARENT,
			OPTION_OUTPUT_FORCE,
			OPTION_INPUT,
			OPTION_INPUT_START_BYTE,
			OPTION_INPUT_START_HUNK,
			OPTION_INPUT_LENGTH_BYTES,
			OPTION_INPUT_LENGTH_HUNKS,
			OPTION_HUNK_SIZE,
			OPTION_COMPRESSION,
			OPTION_IDENT,
			OPTION_CHS,
			OPTION_SECTOR_SIZE,
			OPTION_NUMPROCESSORS
		}
	},

	{ COMMAND_CREATE_CD, do_create_cd, ": create a CD CHD from the input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_PARENT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_HUNK_SIZE,
			OPTION_COMPRESSION,
			OPTION_NUMPROCESSORS
		}
	},

	{ COMMAND_CREATE_LD, do_create_ld, ": create a laserdisc CHD from the input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_PARENT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_START_FRAME,
			OPTION_INPUT_LENGTH_FRAMES,
			OPTION_HUNK_SIZE,
			OPTION_COMPRESSION,
			OPTION_NUMPROCESSORS
		}
	},

	{ COMMAND_EXTRACT_RAW, do_extract_raw, ": extract raw file from a CHD input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_PARENT,
			OPTION_INPUT_START_BYTE,
			OPTION_INPUT_START_HUNK,
			OPTION_INPUT_LENGTH_BYTES,
			OPTION_INPUT_LENGTH_HUNKS
		}
	},

	{ COMMAND_EXTRACT_HD, do_extract_raw, ": extract raw hard disk file from a CHD input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_PARENT,
			OPTION_INPUT_START_BYTE,
			OPTION_INPUT_START_HUNK,
			OPTION_INPUT_LENGTH_BYTES,
			OPTION_INPUT_LENGTH_HUNKS
		}
	},

	{ COMMAND_EXTRACT_CD, do_extract_cd, ": extract CD file from a CHD input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_BIN,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_PARENT,
		}
	},

	{ COMMAND_EXTRACT_LD, do_extract_ld, ": extract laserdisc AVI from a CHD input file",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_PARENT,
			OPTION_INPUT_START_FRAME,
			OPTION_INPUT_LENGTH_FRAMES
		}
	},

	{ COMMAND_COPY, do_copy, ": copy data from one CHD to another of the same type",
		{
			REQUIRED OPTION_OUTPUT,
			OPTION_OUTPUT_PARENT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_INPUT,
			OPTION_INPUT_PARENT,
			OPTION_INPUT_START_BYTE,
			OPTION_INPUT_START_HUNK,
			OPTION_INPUT_LENGTH_BYTES,
			OPTION_INPUT_LENGTH_HUNKS,
			OPTION_HUNK_SIZE,
			OPTION_COMPRESSION,
			OPTION_NUMPROCESSORS
		}
	},

	{ COMMAND_ADD_METADATA, do_add_metadata, ": add metadata to the CHD",
		{
			REQUIRED OPTION_INPUT,
			REQUIRED OPTION_TAG,
			OPTION_INDEX,
			OPTION_VALUE_TEXT,
			OPTION_VALUE_FILE,
			OPTION_NO_CHECKSUM
		}
	},

	{ COMMAND_DEL_METADATA, do_del_metadata, ": remove metadata from the CHD",
		{
			REQUIRED OPTION_INPUT,
			REQUIRED OPTION_TAG,
			OPTION_INDEX
		}
	},

	{ COMMAND_DUMP_METADATA, do_dump_metadata, ": dump metadata from the CHD to stdout or to a file",
		{
			REQUIRED OPTION_INPUT,
			OPTION_OUTPUT,
			OPTION_OUTPUT_FORCE,
			REQUIRED OPTION_TAG,
			OPTION_INDEX
		}
	}
};



//**************************************************************************
//  IMPLEMENTATION
//**************************************************************************

//-------------------------------------------------
//  report_error - report an error
//-------------------------------------------------

static void report_error(int error, const char *format, ...)
{
	// output to stderr
	va_list arg;
	va_start(arg, format);
	vfprintf(stderr, format, arg);
	fflush(stderr);
	va_end(arg);
	fprintf(stderr, "\n");

	// reset time for progress and return the error
	lastprogress = 0;
	throw fatal_error(error);
}


//-------------------------------------------------
//  progress - generic progress callback
//-------------------------------------------------

static void ATTR_PRINTF(2,3) progress(bool forceit, const char *format, ...)
{
	// skip if it hasn't been long enough
	clock_t curtime = clock();
	if (!forceit && lastprogress != 0 && curtime - lastprogress < CLOCKS_PER_SEC / 2)
		return;
	lastprogress = curtime;

	// standard vfprintf stuff here
	va_list arg;
	va_start(arg, format);
	vfprintf(stderr, format, arg);
	fflush(stderr);
	va_end(arg);
}


//-------------------------------------------------
//  print_help - print help for all the commands
//-------------------------------------------------

static int print_help(const char *argv0, const char *error = NULL)
{
	// print the error message first
	if (error != NULL)
		fprintf(stderr, "Error: %s\n\n", error);

	// print a summary of each command
	printf("Usage:\n");
	for (int cmdnum = 0; cmdnum < ARRAY_LENGTH(s_commands); cmdnum++)
	{
		const command_description &desc = s_commands[cmdnum];
		printf("   %s %s%s\n", argv0, desc.name, desc.description);
	}
	printf("\nFor help with any command, run:\n");
	printf("   %s %s <command>\n", argv0, COMMAND_HELP);
	return 1;
}


//-------------------------------------------------
//  print_help - print help for all a specific
//  command
//-------------------------------------------------

static int print_help(const char *argv0, const command_description &desc, const char *error = NULL)
{
	// print the error message first
	if (error != NULL)
		fprintf(stderr, "Error: %s\n\n", error);

	// print usage for this command
	printf("Usage:\n");
	printf("   %s %s [options], where valid options are:\n", argv0, desc.name);
	for (int valid = 0; valid < ARRAY_LENGTH(desc.valid_options); valid++)
	{
		// determine whether we are required
		const char *option = desc.valid_options[valid];
		if (option == NULL)
			break;
		bool required = (option[0] == REQUIRED[0]);
		if (required)
			option++;

		// find the option
		for (int optnum = 0; optnum < ARRAY_LENGTH(s_options); optnum++)
			if (strcmp(option, s_options[optnum].name) == 0)
			{
				const option_description &odesc = s_options[optnum];
				printf("      --%s", odesc.name);
				if (odesc.shortname != NULL)
					printf(", -%s", odesc.shortname);
				printf("%s%s\n", odesc.description, required ? " (required)" : "");
			}
	}
	return 1;
}


//-------------------------------------------------
//  big_int_string - create a 64-bit string
//-------------------------------------------------

const char *big_int_string(astring &string, UINT64 intvalue)
{
	// 0 is a special case
	if (intvalue == 0)
		return string.cpy("0");

	// loop until all chunks are done
	string.reset();
	bool first = true;
	while (intvalue != 0)
	{
		int chunk = intvalue % 1000;
		intvalue /= 1000;

		astring insert;
		insert.format((intvalue != 0) ? "%03d" : "%d", chunk);

		if (!first)
			string.ins(0, ",");
		first = false;
		string.ins(0, insert);
	}
	return string;
}


//-------------------------------------------------
//  msf_string_from_frames - output the given
//  number of frames in M:S:F format
//-------------------------------------------------

const char *msf_string_from_frames(astring &string, UINT32 frames)
{
	return string.format("%02d:%02d:%02d", frames / (75 * 60), (frames / 75) % 60, frames % 75);
}


//-------------------------------------------------
//  parse_number - parse a number string with an
//  optional k/m/g suffix
//-------------------------------------------------

UINT64 parse_number(const char *string)
{
	// 0-length string is 0
	int length = strlen(string);
	if (length == 0)
		return 0;

	// scan forward over digits
	UINT64 result = 0;
	while (isdigit(*string))
	{
		result = (result * 10) + (*string - '0');
		string++;
	}

	// handle multipliers
	if (*string == 'k' || *string == 'K')
		result *= 1024;
	if (*string == 'm' || *string == 'M')
		result *= 1024 * 1024;
	if (*string == 'g' || *string == 'G')
		result *= 1024 * 1024 * 1024;

	return result;
}


//-------------------------------------------------
//  guess_chs - given a file and an offset,
//  compute a best guess CHS value set
//-------------------------------------------------

static void guess_chs(astring *filename, UINT64 filesize, int sectorsize, UINT32 &cylinders, UINT32 &heads, UINT32 &sectors, UINT32 &bps)
{
	// if this is a direct physical drive read, handle it specially
	if (filename != NULL && osd_get_physical_drive_geometry(*filename, &cylinders, &heads, &sectors, &bps))
		return;

	// if we have no length to work with, we can't guess
	if (filesize == 0)
		report_error(1, "Can't guess CHS values because there is no input file");

	// validate the size
	if (filesize % sectorsize != 0)
		report_error(1, "Can't guess CHS values because data size is not divisible by %d", sectorsize);
	;

	// now find a valid value
	for (UINT32 totalsectors = filesize / sectorsize; ; totalsectors++)
		for (UINT32 cursectors = 63; cursectors > 1; cursectors--)
			if (totalsectors % cursectors == 0)
			{
				UINT32 totalheads = totalsectors / cursectors;
				for (UINT32 curheads = 16; curheads > 1; curheads--)
					if (totalheads % curheads == 0)
					{
						cylinders = totalheads / curheads;
						heads = curheads;
						sectors = cursectors;
						return;
					}
			}

	// ack, it didn't work!
	report_error(1, "Can't guess CHS values because no logical combination works!");
}


//-------------------------------------------------
//  parse_input_chd_parameters - parse the
//  standard set of input CHD parameters
//-------------------------------------------------

static void parse_input_chd_parameters(const parameters_t &params, chd_file &input_chd, chd_file &input_parent_chd, bool writeable = false)
{
	// process input parent file
	astring *input_chd_parent_str = params.find(OPTION_INPUT_PARENT);
	if (input_chd_parent_str != NULL)
	{
		chd_error err = input_parent_chd.open(*input_chd_parent_str);
		if (err != CHDERR_NONE)
			report_error(1, "Error opening parent CHD file (%s): %s", input_chd_parent_str->cstr(), chd_file::error_string(err));
	}

	// process input file
	astring *input_chd_str = params.find(OPTION_INPUT);
	if (input_chd_str != NULL)
	{
		chd_error err = input_chd.open(*input_chd_str, writeable, input_parent_chd.opened() ? &input_parent_chd : NULL);
		if (err != CHDERR_NONE)
			report_error(1, "Error opening CHD file (%s): %s", input_chd_str->cstr(), chd_file::error_string(err));
	}
}


//-------------------------------------------------
//  parse_input_start_end - parse input start/end
//  parameters in a standard way
//-------------------------------------------------

static void parse_input_start_end(const parameters_t &params, UINT64 logical_size, UINT32 hunkbytes, UINT32 framebytes, UINT64 &input_start, UINT64 &input_end)
{
	// process start/end if we were provided an input CHD
	input_start = 0;
	input_end = logical_size;

	// process input start
	astring *input_start_byte_str = params.find(OPTION_INPUT_START_BYTE);
	astring *input_start_hunk_str = params.find(OPTION_INPUT_START_HUNK);
	astring *input_start_frame_str = params.find(OPTION_INPUT_START_FRAME);
	if (input_start_byte_str != NULL)
		input_start = parse_number(*input_start_byte_str);
	if (input_start_hunk_str != NULL)
		input_start = parse_number(*input_start_hunk_str) * hunkbytes;
	if (input_start_frame_str != NULL)
		input_start = parse_number(*input_start_frame_str) * framebytes;
	if (input_start >= input_end)
		report_error(1, "Input start offset greater than input file size");

	// process input length
	astring *input_length_bytes_str = params.find(OPTION_INPUT_LENGTH_BYTES);
	astring *input_length_hunks_str = params.find(OPTION_INPUT_LENGTH_HUNKS);
	astring *input_length_frames_str = params.find(OPTION_INPUT_LENGTH_FRAMES);
	UINT64 input_length = input_end;
	if (input_length_bytes_str != NULL)
		input_length = parse_number(*input_length_bytes_str);
	if (input_length_hunks_str != NULL)
		input_length = parse_number(*input_length_hunks_str) * hunkbytes;
	if (input_length_frames_str != NULL)
		input_length = parse_number(*input_length_frames_str) * framebytes;
	if (input_start + input_length < input_end)
		input_end = input_start + input_length;
}


//-------------------------------------------------
//  check_existing_output_file - see if an output
//  file already exists, and error if it does,
//  unless --force is specified
//-------------------------------------------------

static void check_existing_output_file(const parameters_t &params, const char *filename)
{
	if (params.find(OPTION_OUTPUT_FORCE) == NULL)
	{
		core_file *file;
		file_error filerr = core_fopen(filename, OPEN_FLAG_READ, &file);
		if (filerr == FILERR_NONE)
		{
			core_fclose(file);
			report_error(1, "Error: file already exists (%s)\nUse --force (or -f) to force overwriting", filename);
		}
	}
}


//-------------------------------------------------
//  parse_output_chd_parameters - parse the
//  standard set of output CHD parameters
//-------------------------------------------------

static astring *parse_output_chd_parameters(const parameters_t &params, chd_file &output_parent_chd)
{
	// process output parent file
	astring *output_chd_parent_str = params.find(OPTION_OUTPUT_PARENT);
	if (output_chd_parent_str != NULL)
	{
		chd_error err = output_parent_chd.open(*output_chd_parent_str);
		if (err != CHDERR_NONE)
			report_error(1, "Error opening parent CHD file (%s): %s", output_chd_parent_str->cstr(), chd_file::error_string(err));
	}

	// process output file
	astring *output_chd_str = params.find(OPTION_OUTPUT);
	if (output_chd_str != NULL)
		check_existing_output_file(params, *output_chd_str);
	return output_chd_str;
}


//-------------------------------------------------
//  parse_hunk_size - parse the hunk_size
//  parameter in a standard way
//-------------------------------------------------

static void parse_hunk_size(const parameters_t &params, UINT32 required_granularity, UINT32 &hunk_size)
{
	astring *hunk_size_str = params.find(OPTION_HUNK_SIZE);
	if (hunk_size_str != NULL)
	{
		hunk_size = parse_number(*hunk_size_str);
		if (hunk_size < 16 || hunk_size > 1024 * 1024)
			report_error(1, "Invalid hunk size");
		if (hunk_size % required_granularity != 0)
			report_error(1, "Hunk size is not an even multiple of %d", required_granularity);
	}
}


//-------------------------------------------------
//  parse_compression - parse a standard
//  compression parameter string
//-------------------------------------------------

static void parse_compression(const parameters_t &params, chd_codec_type compression[4])
{
	// see if anything was specified
	astring *compression_str = params.find(OPTION_COMPRESSION);
	if (compression_str == NULL)
		return;

	// special case: 'none'
	if (*compression_str == "none")
	{
		compression[0] = compression[1] = compression[2] = compression[3] = CHD_CODEC_NONE;
		return;
	}

	// iterate through compressors
	int index = 0;
	for (int start = 0, end = compression_str->chr(0, ','); index < 4; start = end + 1, end = compression_str->chr(end + 1, ','))
	{
		astring name(*compression_str, start, (end == -1) ? -1 : end - start);
		if (name.len() != 4)
			report_error(1, "Invalid compressor '%s' specified", name.cstr());
		chd_codec_type type = CHD_MAKE_TAG(name[0], name[1], name[2], name[3]);
		if (!chd_codec_list::codec_exists(type))
			report_error(1, "Invalid compressor '%s' specified", name.cstr());
		compression[index++] = type;
		if (end == -1)
			break;
	}
}


//-------------------------------------------------
//  parse_numprocessors - handle the numprocessors
//  command
//-------------------------------------------------

static void parse_numprocessors(const parameters_t &params)
{
	astring *numprocessors_str = params.find(OPTION_NUMPROCESSORS);
	if (numprocessors_str == NULL)
		return;

	int count = atoi(*numprocessors_str);
	if (count > 0)
	{
		extern int osd_num_processors;
		osd_num_processors = count;
	}
}


//-------------------------------------------------
//  compression_string - create a friendly string
//  describing a set of compressors
//-------------------------------------------------

static const char *compression_string(astring &string, chd_codec_type compression[4])
{
	// output compression types
	string.reset();
	if (compression[0] == CHD_CODEC_NONE)
		return string.cpy("none");

	// iterate over types
	for (int index = 0; index < 4; index++)
	{
		chd_codec_type type = compression[index];
		if (type == CHD_CODEC_NONE)
			break;
		if (index != 0)
			string.cat(", ");
		string.cat((type >> 24) & 0xff).cat((type >> 16) & 0xff).cat((type >> 8) & 0xff).cat(type & 0xff);
		string.cat(" (").cat(chd_codec_list::codec_name(type)).cat(")");
	}
	return string;
}


//-------------------------------------------------
//  compress_common - standard compression loop
//-------------------------------------------------

static void compress_common(chd_file_compressor &chd)
{
	// begin compressing
	chd.compress_begin();

	// loop until done
	double complete, ratio;
	chd_error err;
	while ((err = chd.compress_continue(complete, ratio)) == CHDERR_WALKING_PARENT || err == CHDERR_COMPRESSING)
		if (err == CHDERR_WALKING_PARENT)
			progress(false, "Examining parent, %.1f%% complete...  \r", 100.0 * complete);
		else
			progress(false, "Compressing, %.1f%% complete... (ratio=%.1f%%)  \r", 100.0 * complete, 100.0 * ratio);

	// handle errors
	if (err != CHDERR_NONE)
		report_error(1, "Error during compression: %-40s", chd_file::error_string(err));

	// final progress update
	progress(true, "Compression complete ... final ratio = %.1f%%            \n", 100.0 * ratio);
}


//-------------------------------------------------
//  output_track_metadata - output track metadata
//  to a CUE file
//-------------------------------------------------

void output_track_metadata(int mode, core_file *file, int tracknum, const cdrom_track_info &info, const char *filename, UINT32 frameoffs, UINT64 discoffs)
{
	if (mode == MODE_GDI)
	{
        int mode = 0, size = 2048;

        switch (info.trktype)
        {
            case CD_TRACK_MODE1:
                mode = 0;
                size = 2048;
                break;

            case CD_TRACK_MODE1_RAW:
                mode = 4;
                size = 2352;
                break;

            case CD_TRACK_MODE2:
                mode = 4;
                size = 2336;
                break;

            case CD_TRACK_MODE2_FORM1:
                mode = 4;
                size = 2048;
                break;

            case CD_TRACK_MODE2_FORM2:
                mode = 4;
                size = 2324;
                break;

            case CD_TRACK_MODE2_FORM_MIX:
                mode = 4;
                size = 2336;
                break;

            case CD_TRACK_MODE2_RAW:
                mode = 4;
                size = 2352;
                break;

            case CD_TRACK_AUDIO:
                mode = 0;
                size = 2352;
                break;
        }
		core_fprintf(file, "%d %d %d %d %s %" I64FMT "d\n", tracknum+1, frameoffs, mode, size, filename, discoffs);
	}
	else if (mode == MODE_CUEBIN)
    {
        // first track specifies the file
        if (tracknum == 0)
            core_fprintf(file, "FILE \"%s\" BINARY\n", filename);

        // determine submode
        astring tempstr;
        switch (info.trktype)
        {
            case CD_TRACK_MODE1:
            case CD_TRACK_MODE1_RAW:
                tempstr.format("MODE1/%04d", info.datasize);
                break;

            case CD_TRACK_MODE2:
            case CD_TRACK_MODE2_FORM1:
            case CD_TRACK_MODE2_FORM2:
            case CD_TRACK_MODE2_FORM_MIX:
            case CD_TRACK_MODE2_RAW:
                tempstr.format("MODE2/%04d", info.datasize);
                break;

            case CD_TRACK_AUDIO:
                tempstr.cpy("AUDIO");
                break;
        }

        // output TRACK entry
        core_fprintf(file, "  TRACK %02d %s\n", tracknum + 1, tempstr.cstr());

        // output PREGAP
        if (info.pregap > 0)
            core_fprintf(file, "    PREGAP %s\n", msf_string_from_frames(tempstr, info.pregap));

        // output track data
        core_fprintf(file, "    INDEX 01 %s\n", msf_string_from_frames(tempstr, frameoffs));

        // output POSTGAP
        if (info.postgap > 0)
            core_fprintf(file, "    POSTGAP %s\n", msf_string_from_frames(tempstr, info.postgap));
    }
    // non-CUE mode
    else if (mode == MODE_NORMAL)
    {
        // header on the first track
        if (tracknum == 0)
            core_fprintf(file, "CD_ROM\n\n\n");
        core_fprintf(file, "// Track %d\n", tracknum + 1);

        // write out the track type
        astring modesubmode;
        if (info.subtype != CD_SUB_NONE)
            modesubmode.format("%s %s", cdrom_get_type_string(info.trktype), cdrom_get_subtype_string(info.subtype));
        else
            modesubmode.format("%s", cdrom_get_type_string(info.trktype));
        core_fprintf(file, "TRACK %s\n", modesubmode.cstr());

        // write out the attributes
        core_fprintf(file, "NO COPY\n");
        if (info.trktype == CD_TRACK_AUDIO)
        {
            core_fprintf(file, "NO PRE_EMPHASIS\n");
            core_fprintf(file, "TWO_CHANNEL_AUDIO\n");
        }

        // output pregap
        astring tempstr;
        if (info.pregap > 0)
            core_fprintf(file, "ZERO %s %s\n", modesubmode.cstr(), msf_string_from_frames(tempstr, info.pregap));

        // all tracks but the first one have a file offset
        if (tracknum > 0)
            core_fprintf(file, "DATAFILE \"%s\" #%d %s // length in bytes: %d\n", filename, UINT32(discoffs), msf_string_from_frames(tempstr, info.frames), info.frames * (info.datasize + info.subsize));
        else
            core_fprintf(file, "DATAFILE \"%s\" %s // length in bytes: %d\n", filename, msf_string_from_frames(tempstr, info.frames), info.frames * (info.datasize + info.subsize));

        // tracks with pregaps get a START marker too
        if (info.pregap > 0)
            core_fprintf(file, "START %s\n", msf_string_from_frames(tempstr, info.pregap));

        core_fprintf(file, "\n\n");
    }
}


//-------------------------------------------------
//  do_info - dump the header information from
//  a drive image
//-------------------------------------------------

static void do_info(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// print filename and version
	astring tempstr;
	printf("Input file:   %s\n", params.find(OPTION_INPUT)->cstr());
	printf("File Version: %d\n", input_chd.version());
	if (input_chd.version() < 3)
		report_error(1, "Unsupported version (%d); use an older chdman to upgrade to version 3 or later", input_chd.version());

	// output cmpression and size information
	chd_codec_type compression[4] = { input_chd.compression(0), input_chd.compression(1), input_chd.compression(2), input_chd.compression(3) };
	printf("Logical size: %s bytes\n", big_int_string(tempstr, input_chd.logical_bytes()));
	printf("Hunk Size:    %s bytes\n", big_int_string(tempstr, input_chd.hunk_bytes()));
	printf("Total Hunks:  %s\n", big_int_string(tempstr, input_chd.hunk_count()));
	printf("Unit Size:    %s bytes\n", big_int_string(tempstr, input_chd.unit_bytes()));
	printf("Total Units:  %s\n", big_int_string(tempstr, input_chd.unit_count()));
	printf("Compression:  %s\n", compression_string(tempstr, compression));
	printf("CHD size:     %s bytes\n", big_int_string(tempstr, core_fsize(input_chd)));
	if (compression[0] != CHD_CODEC_NONE)
		printf("Ratio:        %.1f%%\n", 100.0 * double(core_fsize(input_chd)) / double(input_chd.logical_bytes()));

	// add SHA1 output
	sha1_t overall = input_chd.sha1();
	if (overall != sha1_t::null)
	{
		printf("SHA1:         %s\n", overall.as_string(tempstr));
		if (input_chd.version() >= 4)
			printf("Data SHA1:    %s\n", input_chd.raw_sha1().as_string(tempstr));
	}
	sha1_t parent = input_chd.parent_sha1();
	if (parent != sha1_t::null)
		printf("Parent SHA1:  %s\n", parent.as_string(tempstr));

	// print out metadata
	dynamic_buffer buffer;
	dynamic_array<metadata_index_info> info;
	for (int index = 0; ; index++)
	{
		// get the indexed metadata item; stop when we hit an error
		chd_metadata_tag metatag;
		UINT8 metaflags;
		chd_error err = input_chd.read_metadata(CHDMETATAG_WILDCARD, index, buffer, metatag, metaflags);
		if (err != CHDERR_NONE)
			break;

		// determine our index
		UINT32 metaindex = ~0;
		for (int cur = 0; cur < info.count(); cur++)
			if (info[cur].tag == metatag)
			{
				metaindex = ++info[cur].index;
				break;
			}

		// if not found, add to our tracking
		if (metaindex == ~0)
		{
			metadata_index_info curinfo = { metatag, 0 };
			info.append(curinfo);
			metaindex = 0;
		}

		// print either a string representation or a hex representation of the tag
		if (isprint((metatag >> 24) & 0xff) && isprint((metatag >> 16) & 0xff) && isprint((metatag >> 8) & 0xff) && isprint(metatag & 0xff))
			printf("Metadata:     Tag='%c%c%c%c'  Index=%d  Length=%d bytes\n", (metatag >> 24) & 0xff, (metatag >> 16) & 0xff, (metatag >> 8) & 0xff, metatag & 0xff, metaindex, buffer.count());
		else
			printf("Metadata:     Tag=%08x  Index=%d  Length=%d bytes\n", metatag, metaindex, buffer.count());
		printf("              ");

		// print up to 60 characters of metadata
		UINT32 count = MIN(60, buffer.count());
		for (int chnum = 0; chnum < count; chnum++)
			printf("%c", isprint(UINT8(buffer[chnum])) ? buffer[chnum] : '.');
		printf("\n");
	}

	// print compression stats if verbose
	if (params.find(OPTION_VERBOSE) != NULL)
	{
		UINT32 compression_types[10] = { 0 };
		for (UINT32 hunknum = 0; hunknum < input_chd.hunk_count(); hunknum++)
		{
			// get info on this hunk
			chd_codec_type codec;
			UINT32 compbytes;
			chd_error err = input_chd.hunk_info(hunknum, codec, compbytes);
			if (err != CHDERR_NONE)
				report_error(1, "Error getting info on hunk %d: %s", hunknum, chd_file::error_string(err));

			// decode into our data
			if (codec > CHD_CODEC_MINI)
				for (int comptype = 0; comptype < 4; comptype++)
					if (codec == input_chd.compression(comptype))
					{
						codec = CHD_CODEC_MINI + 1 + comptype;
						break;
					}
			if (codec > ARRAY_LENGTH(compression_types))
				codec = ARRAY_LENGTH(compression_types) - 1;

			// count stats
			compression_types[codec]++;
		}

		// output the stats
		printf("\n");
		printf("     Hunks  Percent  Name\n");
		printf("----------  -------  ------------------------------------\n");
		for (int comptype = 0; comptype < ARRAY_LENGTH(compression_types); comptype++)
			if (compression_types[comptype] != 0)
			{
				// determine the name
				const char *name = "Unknown";
				switch (comptype)
				{
					case CHD_CODEC_NONE:		name = "Uncompressed";					break;
					case CHD_CODEC_SELF:		name = "Copy from self";				break;
					case CHD_CODEC_PARENT:		name = "Copy from parent";				break;
					case CHD_CODEC_MINI:		name = "Legacy 8-byte mini";			break;
					default:
						int index = comptype - 1 - CHD_CODEC_MINI;
						if (index < 4)
							name = chd_codec_list::codec_name(input_chd.compression(index));
						break;
				}

				// output the stats
				astring tempstr;
				printf("%10s   %5.1f%%  %-40s\n",
						big_int_string(tempstr, compression_types[comptype]),
						100.0 * double(compression_types[comptype]) / double(input_chd.hunk_count()),
						name);
			}
	}
}


//-------------------------------------------------
//  do_verify - validate the SHA1 on a CHD
//-------------------------------------------------

static void do_verify(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// only makes sense for compressed CHDs with valid SHA1's
	if (!input_chd.compressed())
		report_error(0, "No verification to be done; CHD is uncompressed");
	sha1_t raw_sha1 = (input_chd.version() <= 3) ? input_chd.sha1() : input_chd.raw_sha1();
	if (raw_sha1 == sha1_t::null)
		report_error(0, "No verification to be done; CHD has no checksum");

	// create an array to read into
	dynamic_buffer buffer((TEMP_BUFFER_SIZE / input_chd.hunk_bytes()) * input_chd.hunk_bytes());

	// read all the data and build up an SHA-1
	sha1_creator rawsha1;
	for (UINT64 offset = 0; offset < input_chd.logical_bytes(); )
	{
		progress(false, "Verifying, %.1f%% complete... \r", 100.0 * double(offset) / double(input_chd.logical_bytes()));

		// determine how much to read
		UINT32 bytes_to_read = MIN(buffer.count(), input_chd.logical_bytes() - offset);
		chd_error err = input_chd.read_bytes(offset, buffer, bytes_to_read);
		if (err != CHDERR_NONE)
			report_error(1, "Error reading CHD file (%s): %s", params.find(OPTION_INPUT)->cstr(), chd_file::error_string(err));

		// add to the checksum
		rawsha1.append(buffer, bytes_to_read);
		offset += bytes_to_read;
	}
	sha1_t computed_sha1 = rawsha1.finish();

	// finish up
	if (raw_sha1 != computed_sha1)
	{
		astring tempstr;
		fprintf(stderr, "Error: Raw SHA1 in header = %s\n", raw_sha1.as_string(tempstr));
		fprintf(stderr, "              actual SHA1 = %s\n", computed_sha1.as_string(tempstr));

		// fix it if requested; this also fixes the overall one so we don't need to do any more
		if (params.find(OPTION_FIX) != NULL)
		{
			input_chd.set_raw_sha1(computed_sha1);
			printf("SHA-1 updated to correct value in input CHD\n");
		}
	}
	else
	{
		printf("Raw SHA1 verification successful!\n");

		// now include the metadata for >= v4
		if (input_chd.version() >= 4)
		{
			sha1_t computed_overall_sha1 = input_chd.compute_overall_sha1(computed_sha1);
			if (input_chd.sha1() == computed_overall_sha1)
				printf("Overall SHA1 verification successful!\n");
			else
			{
				astring tempstr;
				fprintf(stderr, "Error: Overall SHA1 in header = %s\n", input_chd.sha1().as_string(tempstr));
				fprintf(stderr, "                  actual SHA1 = %s\n", computed_overall_sha1.as_string(tempstr));

				// fix it if requested
				if (params.find(OPTION_FIX) != NULL)
				{
					input_chd.set_raw_sha1(computed_sha1);
					printf("SHA-1 updated to correct value in input CHD\n");
				}
			}
		}
	}
}


//-------------------------------------------------
//  do_create_raw - create a new compressed raw
//  image from a raw file
//-------------------------------------------------

static void do_create_raw(parameters_t &params)
{
	// process input file
	core_file *input_file = NULL;
	astring *input_file_str = params.find(OPTION_INPUT);
	if (input_file_str != NULL)
	{
		file_error filerr = core_fopen(*input_file_str, OPEN_FLAG_READ, &input_file);
		if (filerr != FILERR_NONE)
			report_error(1, "Unable to open file (%s)", input_file_str->cstr());
	}

	// process output CHD
	chd_file output_parent;
	astring *output_chd_str = parse_output_chd_parameters(params, output_parent);

	// process hunk size
	UINT32 hunk_size = output_parent.opened() ? output_parent.hunk_bytes() : 0;
	parse_hunk_size(params, 1, hunk_size);

	// process unit size
	UINT32 unit_size = output_parent.opened() ? output_parent.unit_bytes() : 0;
	astring *unit_size_str = params.find(OPTION_UNIT_SIZE);
	if (unit_size_str != NULL)
	{
		unit_size = parse_number(*unit_size_str);
		if (hunk_size % unit_size != 0)
			report_error(1, "Unit size is not an even divisor of the hunk size");
	}

	// process input start/end (needs to know hunk_size)
	UINT64 input_start;
	UINT64 input_end;
	parse_input_start_end(params, core_fsize(input_file), hunk_size, hunk_size, input_start, input_end);

	// process compression
	chd_codec_type compression[4];
	memcpy(compression, s_default_raw_compression, sizeof(compression));
	parse_compression(params, compression);

	// process numprocessors
	parse_numprocessors(params);

	// print some info
	astring tempstr;
	printf("Output CHD:   %s\n", output_chd_str->cstr());
	if (output_parent.opened())
		printf("Parent CHD:   %s\n", params.find(OPTION_OUTPUT_PARENT)->cstr());
	printf("Input file:   %s\n", input_file_str->cstr());
	if (input_start != 0 || input_end != core_fsize(input_file))
	{
		printf("Input start:  %s\n", big_int_string(tempstr, input_start));
		printf("Input length: %s\n", big_int_string(tempstr, input_end - input_start));
	}
	printf("Compression:  %s\n", compression_string(tempstr, compression));
	printf("Hunk size:    %s\n", big_int_string(tempstr, hunk_size));
	printf("Logical size: %s\n", big_int_string(tempstr, input_end - input_start));

	// catch errors so we can close & delete the output file
	chd_rawfile_compressor *chd = NULL;
	try
	{
		// create the new CHD
		chd = new chd_rawfile_compressor(input_file, input_start, input_end);
		chd_error err;
		if (output_parent.opened())
			err = chd->create(*output_chd_str, input_end - input_start, hunk_size, compression, output_parent);
		else
			err = chd->create(*output_chd_str, input_end - input_start, hunk_size, unit_size, compression);
		if (err != CHDERR_NONE)
			report_error(1, "Error creating CHD file (%s): %s", output_chd_str->cstr(), chd_file::error_string(err));

		// if we have a parent, copy forward all the metadata
		if (output_parent.opened())
			chd->clone_all_metadata(output_parent);

		// compress it generically
		compress_common(*chd);
		delete chd;
	}
	catch (...)
	{
		// delete the output file
		astring *output_chd_str = params.find(OPTION_OUTPUT);
		if (output_chd_str != NULL)
			osd_rmfile(*output_chd_str);
		delete chd;
		throw;
	}
}


//-------------------------------------------------
//  do_create_hd - create a new compressed hard
//  disk image from a raw file
//-------------------------------------------------

static void do_create_hd(parameters_t &params)
{
	// process input file
	core_file *input_file = NULL;
	astring *input_file_str = params.find(OPTION_INPUT);
	if (input_file_str != NULL)
	{
		file_error filerr = core_fopen(*input_file_str, OPEN_FLAG_READ, &input_file);
		if (filerr != FILERR_NONE)
			report_error(1, "Unable to open file (%s)", input_file_str->cstr());
	}

	// process output CHD
	chd_file output_parent;
	astring *output_chd_str = parse_output_chd_parameters(params, output_parent);

	// process sectorsize
	UINT32 sector_size = output_parent.opened() ? output_parent.unit_bytes() : IDE_SECTOR_SIZE;
	astring *sectorsize_str = params.find(OPTION_SECTOR_SIZE);
	if (sectorsize_str != NULL)
	{
		if (output_parent.opened())
			report_error(1, "Sector size does not apply when creating a diff from the parent");
		sector_size = parse_number(*sectorsize_str);
	}

	// process hunk size (needs to know sector_size)
	UINT32 hunk_size = output_parent.opened() ? output_parent.hunk_bytes() : MAX((4096 / sector_size) * sector_size, sector_size);
	parse_hunk_size(params, sector_size, hunk_size);

	// process input start/end (needs to know hunk_size)
	UINT64 input_start = 0;
	UINT64 input_end = 0;
	if (input_file != NULL)
		parse_input_start_end(params, core_fsize(input_file), hunk_size, hunk_size, input_start, input_end);

	// process compression
	chd_codec_type compression[4];
	memcpy(compression, s_default_hd_compression, sizeof(compression));
	if (input_file == NULL)
		compression[0] = compression[1] = compression[2] = compression[3] = CHD_CODEC_NONE;
	parse_compression(params, compression);
	if (input_file == NULL && compression[0] != CHD_CODEC_NONE)
		report_error(1, "Blank hard disks must be uncompressed");

	// process numprocessors
	parse_numprocessors(params);

	// process chs
	UINT32 cylinders = 0;
	UINT32 heads = 0;
	UINT32 sectors = 0;
	astring *chs_str = params.find(OPTION_CHS);
	if (chs_str != NULL)
	{
		if (output_parent.opened())
			report_error(1, "CHS does not apply when creating a diff from the parent");
		if (sscanf(*chs_str, "%d,%d,%d", &cylinders, &heads, &sectors) != 3)
			report_error(1, "Invalid CHS string; must be of the form <cylinders>,<heads>,<sectors>");
	}

	// process ident
	dynamic_buffer identdata;
	if (output_parent.opened())
		output_parent.read_metadata(HARD_DISK_IDENT_METADATA_TAG, 0, identdata);
	astring *ident_str = params.find(OPTION_IDENT);
	if (ident_str != NULL)
	{
		// load the file
		file_error filerr = core_fload(*ident_str, identdata);
		if (filerr != FILERR_NONE)
			report_error(1, "Error reading ident file (%s)", ident_str->cstr());

		// must be at least 14 bytes; extract CHS data from there
		if (identdata.count() < 14)
			report_error(1, "Ident file '%s' is invalid (too short)", ident_str->cstr());
		cylinders = (identdata[3] << 8) | identdata[2];
		heads = (identdata[7] << 8) | identdata[6];
		sectors = (identdata[13] << 8) | identdata[12];
	}

	// extract geometry from the parent if we have one
	if (output_parent.opened() && cylinders == 0)
	{
		astring metadata;
		if (output_parent.read_metadata(HARD_DISK_METADATA_TAG, 0, metadata) != CHDERR_NONE)
			report_error(1, "Unable to find hard disk metadata in parent CHD");
		if (sscanf(metadata, HARD_DISK_METADATA_FORMAT, &cylinders, &heads, &sectors, &sector_size) != 4)
			report_error(1, "Error parsing hard disk metadata in parent CHD");
	}

	// if no CHS values, try to guess them
	if (cylinders == 0)
	{
		if (input_file == NULL && input_end - input_start == 0)
			report_error(1, "Blank hard drives must specify either a length or a set of CHS values");
		guess_chs(input_file_str, input_end - input_start, sector_size, cylinders, heads, sectors, sector_size);
	}
	UINT32 totalsectors = cylinders * heads * sectors;

	// print some info
	astring tempstr;
	printf("Output CHD:   %s\n", output_chd_str->cstr());
	if (output_parent.opened())
		printf("Parent CHD:   %s\n", params.find(OPTION_OUTPUT_PARENT)->cstr());
	if (input_file != NULL)
	{
		printf("Input file:   %s\n", input_file_str->cstr());
		if (input_start != 0 || input_end != core_fsize(input_file))
		{
			printf("Input start:  %s\n", big_int_string(tempstr, input_start));
			printf("Input length: %s\n", big_int_string(tempstr, input_end - input_start));
		}
	}
	printf("Compression:  %s\n", compression_string(tempstr, compression));
	printf("Cylinders:    %d\n", cylinders);
	printf("Heads:        %d\n", heads);
	printf("Sectors:      %d\n", sectors);
	printf("Bytes/sector: %d\n", sector_size);
	printf("Sectors/hunk: %d\n", hunk_size / sector_size);
	printf("Logical size: %s\n", big_int_string(tempstr, UINT64(totalsectors) * UINT64(sector_size)));

	// catch errors so we can close & delete the output file
	chd_rawfile_compressor *chd = NULL;
	try
	{
		// create the new hard drive
		chd = new chd_rawfile_compressor(input_file, input_start, input_end);
		chd_error err;
		if (output_parent.opened())
			err = chd->create(*output_chd_str, UINT64(totalsectors) * UINT64(sector_size), hunk_size, compression, output_parent);
		else
			err = chd->create(*output_chd_str, UINT64(totalsectors) * UINT64(sector_size), hunk_size, sector_size, compression);
		if (err != CHDERR_NONE)
			report_error(1, "Error creating CHD file (%s): %s", output_chd_str->cstr(), chd_file::error_string(err));

		// add the standard hard disk metadata
		astring metadata;
		metadata.format(HARD_DISK_METADATA_FORMAT, cylinders, heads, sectors, sector_size);
		err = chd->write_metadata(HARD_DISK_METADATA_TAG, 0, metadata);
		if (err != CHDERR_NONE)
			report_error(1, "Error adding hard disk metadata: %s", chd_file::error_string(err));

		// write the ident if present
		if (identdata.count() > 0)
		{
			err = chd->write_metadata(HARD_DISK_IDENT_METADATA_TAG, 0, identdata);
			if (err != CHDERR_NONE)
				report_error(1, "Error adding hard disk metadata: %s", chd_file::error_string(err));
		}

		// compress it generically
		if (input_file != NULL)
			compress_common(*chd);
		delete chd;
	}
	catch (...)
	{
		// delete the output file
		astring *output_chd_str = params.find(OPTION_OUTPUT);
		if (output_chd_str != NULL)
			osd_rmfile(*output_chd_str);
		delete chd;
		throw;
	}
}


//-------------------------------------------------
//  do_create_cd - create a new compressed CD
//  image from a raw file
//-------------------------------------------------

static void do_create_cd(parameters_t &params)
{
	// process input file
	chdcd_track_input_info track_info;
	cdrom_toc toc = { 0 };
	astring *input_file_str = params.find(OPTION_INPUT);
	if (input_file_str != NULL)
	{
		chd_error err = chdcd_parse_toc(*input_file_str, toc, track_info);
		if (err != CHDERR_NONE)
			report_error(1, "Error parsing input file (%s: %s)\n", input_file_str->cstr(), chd_file::error_string(err));
	}

	// process output CHD
	chd_file output_parent;
	astring *output_chd_str = parse_output_chd_parameters(params, output_parent);

	// process hunk size
	UINT32 hunk_size = output_parent.opened() ? output_parent.hunk_bytes() : CD_FRAMES_PER_HUNK * CD_FRAME_SIZE;
	parse_hunk_size(params, CD_FRAME_SIZE, hunk_size);

	// process compression
	chd_codec_type compression[4];
	memcpy(compression, s_default_cd_compression, sizeof(compression));
	parse_compression(params, compression);

	// process numprocessors
	parse_numprocessors(params);

	// pad each track to a 4-frame boundry. cdrom.c will deal with this on the read side
	UINT32 origtotalsectors = 0;
	UINT32 totalsectors = 0;
	for (int tracknum = 0; tracknum < toc.numtrks; tracknum++)
	{
		cdrom_track_info &trackinfo = toc.tracks[tracknum];
		int padded = (trackinfo.frames + CD_TRACK_PADDING - 1) / CD_TRACK_PADDING;
		trackinfo.extraframes = padded * CD_TRACK_PADDING - trackinfo.frames;
		origtotalsectors += trackinfo.frames;
		totalsectors += trackinfo.frames + trackinfo.extraframes;
	}

	// print some info
	astring tempstr;
	printf("Output CHD:   %s\n", output_chd_str->cstr());
	if (output_parent.opened())
		printf("Parent CHD:   %s\n", params.find(OPTION_OUTPUT_PARENT)->cstr());
	printf("Input file:   %s\n", input_file_str->cstr());
	printf("Input tracks: %d\n", toc.numtrks);
	printf("Input length: %s\n", msf_string_from_frames(tempstr, origtotalsectors));
	printf("Compression:  %s\n", compression_string(tempstr, compression));
	printf("Logical size: %s\n", big_int_string(tempstr, UINT64(totalsectors) * CD_FRAME_SIZE));

	// catch errors so we can close & delete the output file
	chd_cd_compressor *chd = NULL;
	try
	{
		// create the new CD
		chd = new chd_cd_compressor(toc, track_info);
		chd_error err;
		if (output_parent.opened())
			err = chd->create(*output_chd_str, UINT64(totalsectors) * UINT64(CD_FRAME_SIZE), hunk_size, compression, output_parent);
		else
			err = chd->create(*output_chd_str, UINT64(totalsectors) * UINT64(CD_FRAME_SIZE), hunk_size, CD_FRAME_SIZE, compression);
		if (err != CHDERR_NONE)
			report_error(1, "Error creating CHD file (%s): %s", output_chd_str->cstr(), chd_file::error_string(err));

		// add the standard CD metadata; we do this even if we have a parent because it might be different
		err = cdrom_write_metadata(chd, &toc);
		if (err != CHDERR_NONE)
			report_error(1, "Error adding CD metadata: %s", chd_file::error_string(err));

		// compress it generically
		compress_common(*chd);
		delete chd;
	}
	catch (...)
	{
		// delete the output file
		astring *output_chd_str = params.find(OPTION_OUTPUT);
		if (output_chd_str != NULL)
			osd_rmfile(*output_chd_str);
		delete chd;
		throw;
	}
}


//-------------------------------------------------
//  do_create_ld - create a new A/V file from an
//  input AVI file and metadata
//-------------------------------------------------

static void do_create_ld(parameters_t &params)
{
	// process input file
	avi_file *input_file = NULL;
	astring *input_file_str = params.find(OPTION_INPUT);
	if (input_file_str != NULL)
	{
		avi_error avierr = avi_open(*input_file_str, &input_file);
		if (avierr != AVIERR_NONE)
			report_error(1, "Error opening AVI file (%s): %s\n", input_file_str->cstr(), avi_error_string(avierr));
	}
	const avi_movie_info *aviinfo = avi_get_movie_info(input_file);

	// process output CHD
	chd_file output_parent;
	astring *output_chd_str = parse_output_chd_parameters(params, output_parent);

	// process input start/end
	UINT64 input_start;
	UINT64 input_end;
	parse_input_start_end(params, aviinfo->video_numsamples, 0, 1, input_start, input_end);

	// determine parameters of the incoming video stream
	avi_info info;
	info.fps_times_1million = UINT64(aviinfo->video_timescale) * 1000000 / aviinfo->video_sampletime;
	info.width = aviinfo->video_width;
	info.height = aviinfo->video_height;
	info.interlaced = ((info.fps_times_1million / 1000000) <= 30) && (info.height % 2 == 0) && (info.height > 288);
	info.channels = aviinfo->audio_channels;
	info.rate = aviinfo->audio_samplerate;

	// adjust for interlacing
	if (info.interlaced)
	{
		info.fps_times_1million *= 2;
		info.height /= 2;
		input_start *= 2;
		input_end *= 2;
	}

	// determine the number of bytes per frame
	info.max_samples_per_frame = (UINT64(info.rate) * 1000000 + info.fps_times_1million - 1) / info.fps_times_1million;
	info.bytes_per_frame = avhuff_encoder::raw_data_size(info.width, info.height, info.channels, info.max_samples_per_frame);

	// process hunk size
	UINT32 hunk_size = output_parent.opened() ? output_parent.hunk_bytes() : info.bytes_per_frame;
	parse_hunk_size(params, info.bytes_per_frame, hunk_size);

	// process compression
	chd_codec_type compression[4];
	memcpy(compression, s_default_ld_compression, sizeof(compression));
	parse_compression(params, compression);

	// process numprocessors
	parse_numprocessors(params);

	// print some info
	astring tempstr;
	printf("Output CHD:   %s\n", output_chd_str->cstr());
	if (output_parent.opened())
		printf("Parent CHD:   %s\n", params.find(OPTION_OUTPUT_PARENT)->cstr());
	printf("Input file:   %s\n", input_file_str->cstr());
	if (input_start != 0 && input_end != aviinfo->video_numsamples)
		printf("Input start:  %s\n", big_int_string(tempstr, input_start));
	printf("Input length: %s (%02d:%02d:%02d)\n", big_int_string(tempstr, input_end - input_start),
			UINT32((UINT64(input_end - input_start) * 1000000 / info.fps_times_1million / 60 / 60)),
			UINT32(((UINT64(input_end - input_start) * 1000000 / info.fps_times_1million / 60) % 60)),
			UINT32(((UINT64(input_end - input_start) * 1000000 / info.fps_times_1million) % 60)));
	printf("Frame rate:   %d.%06d\n", info.fps_times_1million / 1000000, info.fps_times_1million % 1000000);
	printf("Frame size:   %d x %d %s\n", info.width, info.height * (info.interlaced ? 2 : 1), info.interlaced ? "interlaced" : "non-interlaced");
	printf("Audio:        %d channels at %d Hz\n", info.channels, info.rate);
	printf("Compression:  %s\n", compression_string(tempstr, compression));
	printf("Hunk size:    %s\n", big_int_string(tempstr, hunk_size));
	printf("Logical size: %s\n", big_int_string(tempstr, UINT64(input_end - input_start) * hunk_size));

	// catch errors so we can close & delete the output file
	chd_avi_compressor *chd = NULL;
	try
	{
		// create the new CHD
		chd = new chd_avi_compressor(*input_file, info, input_start, input_end);
		chd_error err;
		if (output_parent.opened())
			err = chd->create(*output_chd_str, UINT64(input_end - input_start) * hunk_size, hunk_size, compression, output_parent);
		else
			err = chd->create(*output_chd_str, UINT64(input_end - input_start) * hunk_size, hunk_size, info.bytes_per_frame, compression);
		if (err != CHDERR_NONE)
			report_error(1, "Error creating CHD file (%s): %s", output_chd_str->cstr(), chd_file::error_string(err));

		// write the core A/V metadata
		astring metadata;
		metadata.format(AV_METADATA_FORMAT, info.fps_times_1million / 1000000, info.fps_times_1million % 1000000, info.width, info.height, info.interlaced, info.channels, info.rate);
		err = chd->write_metadata(AV_METADATA_TAG, 0, metadata);
		if (err != CHDERR_NONE)
			report_error(1, "Error adding AV metadata: %s\n", chd_file::error_string(err));

		// create the compressor and then run it generically
		compress_common(*chd);

		// write the final LD metadata
		if (info.height == 524/2 || info.height == 624/2)
		{
			err = chd->write_metadata(AV_LD_METADATA_TAG, 0, chd->ldframedata(), 0);
			if (err != CHDERR_NONE)
				report_error(1, "Error adding AVLD metadata: %s\n", chd_file::error_string(err));
		}
		delete chd;
	}
	catch (...)
	{
		// delete the output file
		astring *output_chd_str = params.find(OPTION_OUTPUT);
		if (output_chd_str != NULL)
			osd_rmfile(*output_chd_str);
		delete chd;
		throw;
	}
}


//-------------------------------------------------
//  do_copy - create a new CHD with data from
//  another CHD
//-------------------------------------------------

static void do_copy(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// parse out input start/end
	UINT64 input_start;
	UINT64 input_end;
	parse_input_start_end(params, input_chd.logical_bytes(), input_chd.hunk_bytes(), input_chd.hunk_bytes(), input_start, input_end);

	// process output CHD
	chd_file output_parent;
	astring *output_chd_str = parse_output_chd_parameters(params, output_parent);

	// process hunk size
	UINT32 hunk_size = input_chd.hunk_bytes();
	parse_hunk_size(params, 1, hunk_size);
	if (hunk_size % input_chd.hunk_bytes() != 0 && input_chd.hunk_bytes() % hunk_size != 0)
		report_error(1, "Hunk size is not an even multiple or divisor of input hunk size");

	// process compression; we default to our current preferences using metadata to pick the type
	chd_codec_type compression[4];
	{
		dynamic_buffer metadata;
		if (input_chd.read_metadata(HARD_DISK_METADATA_TAG, 0, metadata) == CHDERR_NONE)
			memcpy(compression, s_default_hd_compression, sizeof(compression));
		else if (input_chd.read_metadata(AV_METADATA_TAG, 0, metadata) == CHDERR_NONE)
			memcpy(compression, s_default_ld_compression, sizeof(compression));
		else if (input_chd.read_metadata(CDROM_OLD_METADATA_TAG, 0, metadata) == CHDERR_NONE ||
				 input_chd.read_metadata(CDROM_TRACK_METADATA_TAG, 0, metadata) == CHDERR_NONE ||
				 input_chd.read_metadata(CDROM_TRACK_METADATA2_TAG, 0, metadata) == CHDERR_NONE)
			memcpy(compression, s_default_cd_compression, sizeof(compression));
		else
			memcpy(compression, s_default_raw_compression, sizeof(compression));
	}
	parse_compression(params, compression);

	// process numprocessors
	parse_numprocessors(params);

	// print some info
	astring tempstr;
	printf("Output CHD:   %s\n", output_chd_str->cstr());
	if (output_parent.opened())
		printf("Parent CHD:   %s\n", params.find(OPTION_OUTPUT_PARENT)->cstr());
	printf("Input CHD:    %s\n", params.find(OPTION_INPUT)->cstr());
	if (input_start != 0 || input_end != input_chd.logical_bytes())
	{
		printf("Input start:  %s\n", big_int_string(tempstr, input_start));
		printf("Input length: %s\n", big_int_string(tempstr, input_end - input_start));
	}
	printf("Compression:  %s\n", compression_string(tempstr, compression));
	printf("Hunk size:    %s\n", big_int_string(tempstr, hunk_size));
	printf("Logical size: %s\n", big_int_string(tempstr, input_end - input_start));

	// catch errors so we can close & delete the output file
	chd_chdfile_compressor *chd = NULL;
	try
	{
		// create the new CHD
		chd = new chd_chdfile_compressor(input_chd, input_start, input_end);
		chd_error err;
		if (output_parent.opened())
			err = chd->create(*output_chd_str, input_end - input_start, hunk_size, compression, output_parent);
		else
			err = chd->create(*output_chd_str, input_end - input_start, hunk_size, input_chd.unit_bytes(), compression);
		if (err != CHDERR_NONE)
			report_error(1, "Error creating CHD file (%s): %s", output_chd_str->cstr(), chd_file::error_string(err));

		// clone all the metadata, upgrading where appropriate
		dynamic_buffer metadata;
		chd_metadata_tag metatag;
		UINT8 metaflags;
		UINT32 index = 0;
		bool redo_cd = false;
		for (err = input_chd.read_metadata(CHDMETATAG_WILDCARD, index++, metadata, metatag, metaflags); err == CHDERR_NONE; err = input_chd.read_metadata(CHDMETATAG_WILDCARD, index++, metadata, metatag, metaflags))
		{
			// if this is an old CD-CHD tag, note that we want to re-do it
			if (metatag == CDROM_OLD_METADATA_TAG || metatag == CDROM_TRACK_METADATA_TAG)
			{
				redo_cd = true;
				continue;
			}

			// otherwise, clone it
			err = chd->write_metadata(metatag, CHDMETAINDEX_APPEND, metadata, metaflags);
			if (err != CHDERR_NONE)
				report_error(1, "Error writing cloned metadata: %s", chd_file::error_string(err));
		}

		// if we need to re-do the CD metadata, do it now
		if (redo_cd)
		{
			cdrom_file *cdrom = cdrom_open(&input_chd);
			if (cdrom == NULL)
				report_error(1, "Error upgrading CD metadata");
			const cdrom_toc *toc = cdrom_get_toc(cdrom);
			err = cdrom_write_metadata(chd, toc);
			if (err != CHDERR_NONE)
				report_error(1, "Error writing upgraded CD metadata: %s", chd_file::error_string(err));
		}

		// compress it generically
		compress_common(*chd);
		delete chd;
	}
	catch (...)
	{
		// delete the output file
		astring *output_chd_str = params.find(OPTION_OUTPUT);
		if (output_chd_str != NULL)
			osd_rmfile(*output_chd_str);
		delete chd;
		throw;
	}
}


//-------------------------------------------------
//  do_extract_raw - extract a raw file from a
//  CHD image
//-------------------------------------------------

static void do_extract_raw(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// parse out input start/end
	UINT64 input_start;
	UINT64 input_end;
	parse_input_start_end(params, input_chd.logical_bytes(), input_chd.hunk_bytes(), input_chd.hunk_bytes(), input_start, input_end);

	// verify output file doesn't exist
	astring *output_file_str = params.find(OPTION_OUTPUT);
	if (output_file_str != NULL)
		check_existing_output_file(params, *output_file_str);

	// print some info
	astring tempstr;
	printf("Output File:  %s\n", output_file_str->cstr());
	printf("Input CHD:    %s\n", params.find(OPTION_INPUT)->cstr());
	if (input_start != 0 || input_end != input_chd.logical_bytes())
	{
		printf("Input start:  %s\n", big_int_string(tempstr, input_start));
		printf("Input length: %s\n", big_int_string(tempstr, input_end - input_start));
	}

	// catch errors so we can close & delete the output file
	core_file *output_file = NULL;
	try
	{
		// process output file
		file_error filerr = core_fopen(*output_file_str, OPEN_FLAG_WRITE | OPEN_FLAG_CREATE, &output_file);
		if (filerr != FILERR_NONE)
			report_error(1, "Unable to open file (%s)", output_file_str->cstr());

		// copy all data
		dynamic_buffer buffer((TEMP_BUFFER_SIZE / input_chd.hunk_bytes()) * input_chd.hunk_bytes());
		for (UINT64 offset = input_start; offset < input_end; )
		{
			progress(false, "Extracting, %.1f%% complete... \r", 100.0 * double(offset - input_start) / double(input_end - input_start));

			// determine how much to read
			UINT32 bytes_to_read = MIN(buffer.count(), input_end - offset);
			chd_error err = input_chd.read_bytes(offset, buffer, bytes_to_read);
			if (err != CHDERR_NONE)
				report_error(1, "Error reading CHD file (%s): %s", params.find(OPTION_INPUT)->cstr(), chd_file::error_string(err));

			// write to the output
			UINT32 count = core_fwrite(output_file, buffer, bytes_to_read);
			if (count != bytes_to_read)
				report_error(1, "Error writing to file; check disk space (%s)", output_file_str->cstr());

			// advance
			offset += bytes_to_read;
		}

		// finish up
		core_fclose(output_file);
		printf("Extraction complete                                    \n");
	}
	catch (...)
	{
		// delete the output file
		if (output_file != NULL)
		{
			core_fclose(output_file);
			osd_rmfile(*output_file_str);
		}
		throw;
	}
}


//-------------------------------------------------
//  do_extract_cd - extract a CD file from a
//  CHD image
//-------------------------------------------------

static void do_extract_cd(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// further process input file
	cdrom_file *cdrom = cdrom_open(&input_chd);
	if (cdrom == NULL)
		report_error(1, "Unable to recognize CHD file as a CD");
	const cdrom_toc *toc = cdrom_get_toc(cdrom);

	// verify output file doesn't exist
	astring *output_file_str = params.find(OPTION_OUTPUT);
	if (output_file_str != NULL)
		check_existing_output_file(params, *output_file_str);

	// verify output BIN file doesn't exist
	astring *output_bin_file_str = params.find(OPTION_OUTPUT_BIN);
	astring default_name(*output_file_str);
	int chop = default_name.rchr(0, '.');
	if (chop != -1)
		default_name.substr(0, chop);
    char basename[128];
    strncpy(basename, default_name.cstr(), 127);
	default_name.cat(".bin");
	if (output_bin_file_str == NULL)
		output_bin_file_str = &default_name;
	check_existing_output_file(params, *output_bin_file_str);

	// print some info
	astring tempstr;
	printf("Output TOC:   %s\n", output_file_str->cstr());
	printf("Output Data:  %s\n", output_bin_file_str->cstr());
	printf("Input CHD:    %s\n", params.find(OPTION_INPUT)->cstr());

	// catch errors so we can close & delete the output file
	core_file *output_bin_file = NULL;
	core_file *output_toc_file = NULL;
	try
	{
        int mode = MODE_NORMAL;

		if (output_file_str->find(".cue") != -1)
        {
            mode = MODE_CUEBIN;
        }
        else if (output_file_str->find(".gdi") != -1)
        {
            mode = MODE_GDI;
        }

        // process output file
        file_error filerr = core_fopen(*output_file_str, OPEN_FLAG_WRITE | OPEN_FLAG_CREATE | OPEN_FLAG_NO_BOM, &output_toc_file);
        if (filerr != FILERR_NONE)
            report_error(1, "Unable to open file (%s)", output_file_str->cstr());

		// process output BIN file
        if (mode != MODE_GDI)
        {
            filerr = core_fopen(*output_bin_file_str, OPEN_FLAG_WRITE | OPEN_FLAG_CREATE, &output_bin_file);
            if (filerr != FILERR_NONE)
                report_error(1, "Unable to open file (%s)", output_bin_file_str->cstr());
        }

		// determine total frames
		UINT64 total_bytes = 0;
		for (int tracknum = 0; tracknum < toc->numtrks; tracknum++)
			total_bytes += toc->tracks[tracknum].frames * (toc->tracks[tracknum].datasize + toc->tracks[tracknum].subsize);

		// GDI must start with the # of tracks
		if (mode == MODE_GDI)
		{
			core_fprintf(output_toc_file, "%d\n", toc->numtrks);
		}

		// iterate over tracks and copy all data
		UINT64 outputoffs = 0;
		UINT32 discoffs = 0;
		dynamic_buffer buffer;
		for (int tracknum = 0; tracknum < toc->numtrks; tracknum++)
		{
            astring trackbin_name(basename);

            if (mode == MODE_GDI)
            {
                char temp[8];
                sprintf(temp, "%02d", tracknum+1);
                trackbin_name.cat(temp);
                trackbin_name.cat(".bin");

                if (output_bin_file)
                {
                    core_fclose(output_bin_file);
                    output_bin_file = NULL;
                }

                filerr = core_fopen(trackbin_name, OPEN_FLAG_WRITE | OPEN_FLAG_CREATE, &output_bin_file);
                if (filerr != FILERR_NONE)
                    report_error(1, "Unable to open file (%s)", trackbin_name.cstr());

                outputoffs = 0;
            }

			// output the metadata about the track to the TOC file
			const cdrom_track_info &trackinfo = toc->tracks[tracknum];
			astring temp;
            if (mode == MODE_GDI)
            {
                output_track_metadata(mode, output_toc_file, tracknum, trackinfo, core_filename_extract_base(temp,trackbin_name), discoffs, outputoffs);
            }
            else
            {
                output_track_metadata(mode, output_toc_file, tracknum, trackinfo, core_filename_extract_base(temp,*output_bin_file_str), discoffs, outputoffs);
            }

            // If this is bin/cue output and the CHD contains subdata, warn the user and don't include
            // the subdata size in the buffer calculation.
			UINT32 output_frame_size = trackinfo.datasize + ((trackinfo.subtype != CD_SUB_NONE) ? trackinfo.subsize : 0);
            if (trackinfo.subtype != CD_SUB_NONE && ((mode == MODE_CUEBIN) || (mode == MODE_GDI)))
            {
                printf("Warning: Track %d has subcode data.  bin/cue and gdi formats cannot contain subcode data and it will be omitted.\n", tracknum+1);
                printf("       : This may affect usage of the output image.  Use bin/toc output to keep all data.\n");
                output_frame_size = trackinfo.datasize;
            }

			// resize the buffer for the track
			buffer.resize((TEMP_BUFFER_SIZE / output_frame_size) * output_frame_size);

			// now read and output the actual data
			UINT32 bufferoffs = 0;
            UINT32 actualframes = trackinfo.frames - trackinfo.padframes;
			for (UINT32 frame = 0; frame < actualframes; frame++)
			{
				progress(false, "Extracting, %.1f%% complete... \r", 100.0 * double(outputoffs) / double(total_bytes));

				// read the data
				cdrom_read_data(cdrom, cdrom_get_track_start(cdrom, tracknum) + frame, &buffer[bufferoffs], trackinfo.trktype);

				// for CDRWin, audio tracks must be reversed
				if ((mode == MODE_CUEBIN) && (trackinfo.trktype == CD_TRACK_AUDIO))
					for (int swapindex = 0; swapindex < trackinfo.datasize; swapindex += 2)
					{
						UINT8 swaptemp = buffer[bufferoffs + swapindex];
						buffer[bufferoffs + swapindex] = buffer[bufferoffs + swapindex + 1];
						buffer[bufferoffs + swapindex + 1] = swaptemp;
					}
				bufferoffs += trackinfo.datasize;
				discoffs++;

				// read the subcode data
				if (trackinfo.subtype != CD_SUB_NONE && (mode == MODE_NORMAL))
				{
                    cdrom_read_subcode(cdrom, cdrom_get_track_start(cdrom, tracknum) + frame, &buffer[bufferoffs]);
                    bufferoffs += trackinfo.subsize;
				}

				// write it out if we need to
				if (bufferoffs == buffer.count() || frame == actualframes - 1)
				{
					core_fseek(output_bin_file, outputoffs, SEEK_SET);
					UINT32 byteswritten = core_fwrite(output_bin_file, buffer, bufferoffs);
					if (byteswritten != bufferoffs)
						report_error(1, "Error writing frame %d to file (%s): %s\n", frame, output_file_str->cstr(), chd_file::error_string(CHDERR_WRITE_ERROR));
					outputoffs += bufferoffs;
					bufferoffs = 0;
				}
			}

            discoffs += trackinfo.padframes;
		}

		// finish up
        core_fclose(output_bin_file);
		core_fclose(output_toc_file);
		printf("Extraction complete                                    \n");
	}
	catch (...)
	{
		// delete the output files
		if (output_bin_file != NULL)
			core_fclose(output_bin_file);
		if (output_toc_file != NULL)
			core_fclose(output_toc_file);
		osd_rmfile(*output_bin_file_str);
		osd_rmfile(*output_file_str);
		throw;
	}
}


//-------------------------------------------------
//  do_extract_ld - extract an AVI file from a
//  CHD image
//-------------------------------------------------

static void do_extract_ld(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// read core metadata
	astring metadata;
	chd_error err = input_chd.read_metadata(AV_METADATA_TAG, 0, metadata);
	if (err != CHDERR_NONE)
		report_error(1, "Unable to find A/V metadata in the input CHD");

	// parse the metadata
	UINT32 fps_times_1million;
	UINT32 max_samples_per_frame;
	UINT32 frame_bytes;
	int width;
	int height;
	int interlaced;
	int channels;
	int rate;
	{
		int fps;
		int fpsfrac;
		if (sscanf(metadata, AV_METADATA_FORMAT, &fps, &fpsfrac, &width, &height, &interlaced, &channels, &rate) != 7)
			report_error(1, "Improperly formatted A/V metadata found");
		fps_times_1million = fps * 1000000 + fpsfrac;
	}
	int interlace_factor = interlaced ? 2 : 1;

	// determine key parameters and validate
	max_samples_per_frame = (UINT64(rate) * 1000000 + fps_times_1million - 1) / fps_times_1million;
	frame_bytes = avhuff_encoder::raw_data_size(width, height, channels, max_samples_per_frame);
	if (frame_bytes != input_chd.hunk_bytes())
		report_error(1, "Frame size does not match hunk size for this CHD");

	// parse out input start/end
	UINT64 input_start;
	UINT64 input_end;
	parse_input_start_end(params, input_chd.hunk_count() / interlace_factor, 0, 1, input_start, input_end);
	input_start *= interlace_factor;
	input_end *= interlace_factor;

	// build up the movie info
	avi_movie_info info;
	info.video_format = FORMAT_YUY2;
	info.video_timescale = fps_times_1million / interlace_factor;
	info.video_sampletime = 1000000;
	info.video_width = width;
	info.video_height = height * interlace_factor;
	info.video_depth = 16;
	info.audio_format = 0;
	info.audio_timescale = rate;
	info.audio_sampletime = 1;
	info.audio_channels = channels;
	info.audio_samplebits = 16;
	info.audio_samplerate = rate;

	// verify output file doesn't exist
	astring *output_file_str = params.find(OPTION_OUTPUT);
	if (output_file_str != NULL)
		check_existing_output_file(params, *output_file_str);

	// print some info
	astring tempstr;
	printf("Output File:  %s\n", output_file_str->cstr());
	printf("Input CHD:    %s\n", params.find(OPTION_INPUT)->cstr());
	if (input_start != 0 || input_end != input_chd.hunk_count())
	{
		printf("Input start:  %s\n", big_int_string(tempstr, input_start));
		printf("Input length: %s\n", big_int_string(tempstr, input_end - input_start));
	}

	// catch errors so we can close & delete the output file
	avi_file *output_file = NULL;
	try
	{
		// process output file
		avi_error avierr = avi_create(*output_file_str, &info, &output_file);
		if (avierr != AVIERR_NONE)
			report_error(1, "Unable to open file (%s)", output_file_str->cstr());

		// create the codec configuration
		avhuff_decompress_config avconfig;
		dynamic_array<INT16> audio_data[16];
		UINT32 actsamples;
		avconfig.maxsamples = max_samples_per_frame;
		avconfig.actsamples = &actsamples;
		for (int chnum = 0; chnum < ARRAY_LENGTH(audio_data); chnum++)
		{
			audio_data[chnum].resize(max_samples_per_frame);
			avconfig.audio[chnum] = audio_data[chnum];
		}

		// iterate over frames
		bitmap_yuy16 fullbitmap(width, height * interlace_factor);
		for (int framenum = input_start; framenum < input_end; framenum++)
		{
			progress(framenum == 0, "Extracting, %.1f%% complete...  \r", 100.0 * double(framenum - input_start) / double(input_end - input_start));

			// set up the fake bitmap for this frame
			avconfig.video.wrap(&fullbitmap.pix(framenum % interlace_factor), fullbitmap.width(), fullbitmap.height() / interlace_factor, fullbitmap.rowpixels() * interlace_factor);
			input_chd.codec_configure(CHD_CODEC_AVHUFF, AVHUFF_CODEC_DECOMPRESS_CONFIG, &avconfig);

			// read the hunk into the buffers
			chd_error err = input_chd.read_hunk(framenum, NULL);
			if (err != CHDERR_NONE)
				report_error(1, "Error reading hunk %d from CHD file (%s): %s\n", framenum, params.find(OPTION_INPUT)->cstr(), chd_file::error_string(err));

			// write audio
			for (int chnum = 0; chnum < channels; chnum++)
			{
				avi_error avierr = avi_append_sound_samples(output_file, chnum, avconfig.audio[chnum], actsamples, 0);
				if (avierr != AVIERR_NONE)
					report_error(1, "Error writing samples for hunk %d to file (%s): %s\n", framenum, output_file_str->cstr(), avi_error_string(avierr));
			}

			// write video
			if (framenum % interlace_factor == interlace_factor - 1)
			{
				avi_error avierr = avi_append_video_frame(output_file, fullbitmap);
				if (avierr != AVIERR_NONE)
					report_error(1, "Error writing video for hunk %d to file (%s): %s\n", framenum, output_file_str->cstr(), avi_error_string(avierr));
			}
		}

		// close and return
		avi_close(output_file);
		printf("Extraction complete                                    \n");
	}
	catch (...)
	{
		// delete the output file
		if (output_file != NULL)
			avi_close(output_file);
		osd_rmfile(*output_file_str);
		throw;
	}
}


//-------------------------------------------------
//  do_add_metadata - add metadata to a CHD from a
//  file
//-------------------------------------------------

static void do_add_metadata(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd, true);

	// process tag
	chd_metadata_tag tag = CHD_MAKE_TAG('?','?','?','?');
	astring *tag_str = params.find(OPTION_TAG);
	if (tag_str != NULL)
	{
		tag_str->cat("    ");
		tag = CHD_MAKE_TAG((*tag_str)[0], (*tag_str)[1], (*tag_str)[2], (*tag_str)[3]);
	}

	// process index
	UINT32 index = 0;
	astring *index_str = params.find(OPTION_INDEX);
	if (index_str != NULL)
		index = atoi(*index_str);

	// process text input
	astring *text_str = params.find(OPTION_VALUE_TEXT);
	astring text;
	if (text_str != NULL)
	{
		text = *text_str;
		if (text[0] == '"' && text[text.len() - 1] == '"')
			text.substr(1, text.len() - 2);
	}

	// process file input
	astring *file_str = params.find(OPTION_VALUE_FILE);
	dynamic_buffer file;
	if (file_str != NULL)
	{
		file_error filerr = core_fload(*file_str, file);
		if (filerr != FILERR_NONE)
			report_error(1, "Error reading metadata file (%s)", file_str->cstr());
	}

	// make sure we have one or the other
	if (text_str == NULL && file_str == NULL)
		report_error(1, "Error: missing either --valuetext/-vt or --valuefile/-vf parameters");
	if (text_str != NULL && file_str != NULL)
		report_error(1, "Error: both --valuetext/-vt or --valuefile/-vf parameters specified; only one permitted");

	// process no checksum
	UINT8 flags = CHD_MDFLAGS_CHECKSUM;
	if (params.find(OPTION_NO_CHECKSUM) != NULL)
		flags &= ~CHD_MDFLAGS_CHECKSUM;

	// print some info
	astring tempstr;
	printf("Input file:   %s\n", params.find(OPTION_INPUT)->cstr());
	printf("Tag:          %c%c%c%c\n", (tag >> 24) & 0xff, (tag >> 16) & 0xff, (tag >> 8) & 0xff, tag & 0xff);
	printf("Index:        %d\n", index);
	if (text_str != NULL)
		printf("Text:         %s\n", text.cstr());
	else
		printf("Data:         %s (%d bytes)\n", file_str->cstr(), file.count());

	// write the metadata
	chd_error err;
	if (text_str != NULL)
		err = input_chd.write_metadata(tag, index, text, flags);
	else
		err = input_chd.write_metadata(tag, index, file, flags);
	if (err != CHDERR_NONE)
		report_error(1, "Error adding metadata: %s", chd_file::error_string(err));
	else
		printf("Metadata added\n");
}


//-------------------------------------------------
//  do_del_metadata - remove metadata from a CHD
//-------------------------------------------------

static void do_del_metadata(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd, true);

	// process tag
	chd_metadata_tag tag = CHD_MAKE_TAG('?','?','?','?');
	astring *tag_str = params.find(OPTION_TAG);
	if (tag_str != NULL)
	{
		tag_str->cat("    ");
		tag = CHD_MAKE_TAG((*tag_str)[0], (*tag_str)[1], (*tag_str)[2], (*tag_str)[3]);
	}

	// process index
	UINT32 index = 0;
	astring *index_str = params.find(OPTION_INDEX);
	if (index_str != NULL)
		index = atoi(*index_str);

	// print some info
	astring tempstr;
	printf("Input file:   %s\n", params.find(OPTION_INPUT)->cstr());
	printf("Tag:          %c%c%c%c\n", (tag >> 24) & 0xff, (tag >> 16) & 0xff, (tag >> 8) & 0xff, tag & 0xff);
	printf("Index:        %d\n", index);

	// write the metadata
	chd_error err = input_chd.delete_metadata(tag, index);
	if (err != CHDERR_NONE)
		report_error(1, "Error removing metadata: %s", chd_file::error_string(err));
	else
		printf("Metadata removed\n");
}


//-------------------------------------------------
//  do_dump_metadata - dump metadata from a CHD
//-------------------------------------------------

static void do_dump_metadata(parameters_t &params)
{
	// parse out input files
	chd_file input_parent_chd;
	chd_file input_chd;
	parse_input_chd_parameters(params, input_chd, input_parent_chd);

	// verify output file doesn't exist
	astring *output_file_str = params.find(OPTION_OUTPUT);
	if (output_file_str != NULL)
		check_existing_output_file(params, *output_file_str);

	// process tag
	chd_metadata_tag tag = CHD_MAKE_TAG('?','?','?','?');
	astring *tag_str = params.find(OPTION_TAG);
	if (tag_str != NULL)
	{
		tag_str->cat("    ");
		tag = CHD_MAKE_TAG((*tag_str)[0], (*tag_str)[1], (*tag_str)[2], (*tag_str)[3]);
	}

	// process index
	UINT32 index = 0;
	astring *index_str = params.find(OPTION_INDEX);
	if (index_str != NULL)
		index = atoi(*index_str);

	// write the metadata
	dynamic_buffer buffer;
	chd_error err = input_chd.read_metadata(tag, index, buffer);
	if (err != CHDERR_NONE)
		report_error(1, "Error reading metadata: %s", chd_file::error_string(err));

	// catch errors so we can close & delete the output file
	core_file *output_file = NULL;
	try
	{
		// create the file
		if (output_file_str != NULL)
		{
			file_error filerr = core_fopen(*output_file_str, OPEN_FLAG_WRITE | OPEN_FLAG_CREATE, &output_file);
			if (filerr != FILERR_NONE)
				report_error(1, "Unable to open file (%s)", output_file_str->cstr());

			// output the metadata
			UINT32 count = core_fwrite(output_file, buffer, buffer.count());
			if (count != buffer.count())
				report_error(1, "Error writing file (%s)", output_file_str->cstr());
			core_fclose(output_file);

			// provide some feedback
			astring tempstr;
			printf("File (%s) written, %s bytes\n", output_file_str->cstr(), big_int_string(tempstr, buffer.count()));
		}

		// flush to stdout
		else
		{
			fwrite(buffer, 1, buffer.count(), stdout);
			fflush(stdout);
		}
	}
	catch (...)
	{
		// delete the output file
		if (output_file != NULL)
			core_fclose(output_file);
		osd_rmfile(*output_file_str);
		throw;
	}
}


//-------------------------------------------------
//  main - entry point
//-------------------------------------------------

int CLIB_DECL main(int argc, char *argv[])
{
	// print the header
	extern const char build_version[];
	printf("chdman - MAME Compressed Hunks of Data (CHD) manager %s\n", build_version);

	// handle help specially
	if (argc < 2)
		return print_help(argv[0]);
	int argnum = 1;
	const char *command = argv[argnum++];
	bool help = (strcmp(command, COMMAND_HELP) == 0);
	if (help)
	{
		if (argc <= 2)
			return print_help(argv[0]);
		command = argv[argnum++];
	}

	// iterate over commands to find our match
	for (int cmdnum = 0; cmdnum < ARRAY_LENGTH(s_commands); cmdnum++)
		if (strcmp(command, s_commands[cmdnum].name) == 0)
		{
			const command_description &desc = s_commands[cmdnum];

			// print help if that was requested
			if (help)
				return print_help(argv[0], desc);

			// otherwise, verify the parameters
			tagmap_t<astring *> parameters;
			while (argnum < argc)
			{
				// should be an option name
				const char *arg = argv[argnum++];
				if (arg[0] != '-')
					return print_help(argv[0], desc, "Expected option, not parameter");

				// iterate over valid options
				int valid;
				for (valid = 0; valid < ARRAY_LENGTH(desc.valid_options); valid++)
				{
					// reduce to the option name
					const char *validname = desc.valid_options[valid];
					if (validname == NULL)
						break;
					if (*validname == REQUIRED[0])
						validname++;

					// find the matching option description
					int optnum;
					for (optnum = 0; optnum < ARRAY_LENGTH(s_options); optnum++)
						if (strcmp(s_options[optnum].name, validname) == 0)
							break;
					assert(optnum != ARRAY_LENGTH(s_options));

					// do we match?
					const option_description &odesc = s_options[optnum];
					if ((arg[1] == '-' && strcmp(odesc.name, &arg[2]) == 0) ||
						(arg[1] != '-' && odesc.shortname != NULL && strcmp(odesc.shortname, &arg[1]) == 0))
					{
						// if we need a parameter, consume it
						const char *param = "";
						if (odesc.parameter)
						{
							if (argnum >= argc || argv[argnum][0] == '-')
								return print_help(argv[0], desc, "Option is missing parameter");
							param = argv[argnum++];
						}

						// add to the map
						if (parameters.add(odesc.name, new astring(param)) == TMERR_DUPLICATE)
							return print_help(argv[0], desc, "Multiple parameters of the same type specified");
						break;
					}
				}

				// if not valid, error
				if (valid == ARRAY_LENGTH(desc.valid_options))
					return print_help(argv[0], desc, "Option not valid for this command");
			}

			// make sure we got all our required parameters
			for (int valid = 0; valid < ARRAY_LENGTH(desc.valid_options); valid++)
			{
				const char *validname = desc.valid_options[valid];
				if (validname == NULL)
					break;
				if (*validname == REQUIRED[0] && parameters.find(++validname) == NULL)
					return print_help(argv[0], desc, "Required parameters missing");
			}

			// all clear, run the command
			try
			{
				(*s_commands[cmdnum].handler)(parameters);
				return 0;
			}
			catch (fatal_error &err)
			{
				return err.error();
			}
			catch (std::bad_alloc &)
			{
				fprintf(stderr, "Out of memory\n");
				return 1;
			}
			catch (...)
			{
				fprintf(stderr, "Unhandled exception\n");
				return 1;
			}
		}

	// print generic help if nothing found
	return print_help(argv[0]);
}