mirrors
/
devilution
1
0
Fork 0
Code Releases Activity

Add files via upload

pull/1/head
galaxyhaxz 2018-06-06 21:22:26 -05:00 committed by GitHub
parent 2668ffe6ac
commit 3fed670473
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 1439 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

522
3rdParty/PKWare/explode.cpp vendored Normal file
View File

@ -0,0 +1,522 @@
/*****************************************************************************/
/* explode.cpp Copyright (c) Ladislav Zezula 2003 */
/*---------------------------------------------------------------------------*/
/* Implode function of PKWARE Data Compression library */
/*---------------------------------------------------------------------------*/
/* Date Ver Who Comment */
/* -------- ---- --- ------- */
/* 11.03.03 1.00 Lad Splitted from Pkware.cpp */
/* 08.04.03 1.01 Lad Renamed to explode.cpp to be compatible with pkware */
/* 02.05.03 1.01 Lad Stress test done */
/* 22.04.10 1.01 Lad Documented */
/*****************************************************************************/
#include <assert.h>
#include <string.h>
#include "pkware.h"
#define PKDCL_OK 0
#define PKDCL_STREAM_END 1 // All data from the input stream is read
#define PKDCL_NEED_DICT 2 // Need more data (dictionary)
#define PKDCL_CONTINUE 10 // Internal flag, not returned to user
#define PKDCL_GET_INPUT 11 // Internal flag, not returned to user
static char CopyrightPkware[] = "PKWARE Data Compression Library for Win32\r\n"
"Copyright 1989-1995 PKWARE Inc. All Rights Reserved\r\n"
"Patent No. 5,051,745\r\n"
"PKWARE Data Compression Library Reg. U.S. Pat. and Tm. Off.\r\n"
"Version 1.11\r\n";
//-----------------------------------------------------------------------------
// Tables
static unsigned char DistBits[] =
{
0x02, 0x04, 0x04, 0x05, 0x05, 0x05, 0x05, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06,
0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08
};
static unsigned char DistCode[] =
{
0x03, 0x0D, 0x05, 0x19, 0x09, 0x11, 0x01, 0x3E, 0x1E, 0x2E, 0x0E, 0x36, 0x16, 0x26, 0x06, 0x3A,
0x1A, 0x2A, 0x0A, 0x32, 0x12, 0x22, 0x42, 0x02, 0x7C, 0x3C, 0x5C, 0x1C, 0x6C, 0x2C, 0x4C, 0x0C,
0x74, 0x34, 0x54, 0x14, 0x64, 0x24, 0x44, 0x04, 0x78, 0x38, 0x58, 0x18, 0x68, 0x28, 0x48, 0x08,
0xF0, 0x70, 0xB0, 0x30, 0xD0, 0x50, 0x90, 0x10, 0xE0, 0x60, 0xA0, 0x20, 0xC0, 0x40, 0x80, 0x00
};
static unsigned char ExLenBits[] =
{
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08
};
static unsigned short LenBase[] =
{
0x0000, 0x0001, 0x0002, 0x0003, 0x0004, 0x0005, 0x0006, 0x0007,
0x0008, 0x000A, 0x000E, 0x0016, 0x0026, 0x0046, 0x0086, 0x0106
};
static unsigned char LenBits[] =
{
0x03, 0x02, 0x03, 0x03, 0x04, 0x04, 0x04, 0x05, 0x05, 0x05, 0x05, 0x06, 0x06, 0x06, 0x07, 0x07
};
static unsigned char LenCode[] =
{
0x05, 0x03, 0x01, 0x06, 0x0A, 0x02, 0x0C, 0x14, 0x04, 0x18, 0x08, 0x30, 0x10, 0x20, 0x40, 0x00
};
static unsigned char ChBitsAsc[] =
{
0x0B, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x08, 0x07, 0x0C, 0x0C, 0x07, 0x0C, 0x0C,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0D, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x04, 0x0A, 0x08, 0x0C, 0x0A, 0x0C, 0x0A, 0x08, 0x07, 0x07, 0x08, 0x09, 0x07, 0x06, 0x07, 0x08,
0x07, 0x06, 0x07, 0x07, 0x07, 0x07, 0x08, 0x07, 0x07, 0x08, 0x08, 0x0C, 0x0B, 0x07, 0x09, 0x0B,
0x0C, 0x06, 0x07, 0x06, 0x06, 0x05, 0x07, 0x08, 0x08, 0x06, 0x0B, 0x09, 0x06, 0x07, 0x06, 0x06,
0x07, 0x0B, 0x06, 0x06, 0x06, 0x07, 0x09, 0x08, 0x09, 0x09, 0x0B, 0x08, 0x0B, 0x09, 0x0C, 0x08,
0x0C, 0x05, 0x06, 0x06, 0x06, 0x05, 0x06, 0x06, 0x06, 0x05, 0x0B, 0x07, 0x05, 0x06, 0x05, 0x05,
0x06, 0x0A, 0x05, 0x05, 0x05, 0x05, 0x08, 0x07, 0x08, 0x08, 0x0A, 0x0B, 0x0B, 0x0C, 0x0C, 0x0C,
0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D,
0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D,
0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x0D, 0x0C, 0x0D, 0x0D, 0x0D, 0x0C, 0x0D, 0x0D, 0x0D, 0x0C, 0x0D, 0x0D, 0x0D, 0x0D, 0x0C, 0x0D,
0x0D, 0x0D, 0x0C, 0x0C, 0x0C, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D
};
static unsigned short ChCodeAsc[] =
{
0x0490, 0x0FE0, 0x07E0, 0x0BE0, 0x03E0, 0x0DE0, 0x05E0, 0x09E0,
0x01E0, 0x00B8, 0x0062, 0x0EE0, 0x06E0, 0x0022, 0x0AE0, 0x02E0,
0x0CE0, 0x04E0, 0x08E0, 0x00E0, 0x0F60, 0x0760, 0x0B60, 0x0360,
0x0D60, 0x0560, 0x1240, 0x0960, 0x0160, 0x0E60, 0x0660, 0x0A60,
0x000F, 0x0250, 0x0038, 0x0260, 0x0050, 0x0C60, 0x0390, 0x00D8,
0x0042, 0x0002, 0x0058, 0x01B0, 0x007C, 0x0029, 0x003C, 0x0098,
0x005C, 0x0009, 0x001C, 0x006C, 0x002C, 0x004C, 0x0018, 0x000C,
0x0074, 0x00E8, 0x0068, 0x0460, 0x0090, 0x0034, 0x00B0, 0x0710,
0x0860, 0x0031, 0x0054, 0x0011, 0x0021, 0x0017, 0x0014, 0x00A8,
0x0028, 0x0001, 0x0310, 0x0130, 0x003E, 0x0064, 0x001E, 0x002E,
0x0024, 0x0510, 0x000E, 0x0036, 0x0016, 0x0044, 0x0030, 0x00C8,
0x01D0, 0x00D0, 0x0110, 0x0048, 0x0610, 0x0150, 0x0060, 0x0088,
0x0FA0, 0x0007, 0x0026, 0x0006, 0x003A, 0x001B, 0x001A, 0x002A,
0x000A, 0x000B, 0x0210, 0x0004, 0x0013, 0x0032, 0x0003, 0x001D,
0x0012, 0x0190, 0x000D, 0x0015, 0x0005, 0x0019, 0x0008, 0x0078,
0x00F0, 0x0070, 0x0290, 0x0410, 0x0010, 0x07A0, 0x0BA0, 0x03A0,
0x0240, 0x1C40, 0x0C40, 0x1440, 0x0440, 0x1840, 0x0840, 0x1040,
0x0040, 0x1F80, 0x0F80, 0x1780, 0x0780, 0x1B80, 0x0B80, 0x1380,
0x0380, 0x1D80, 0x0D80, 0x1580, 0x0580, 0x1980, 0x0980, 0x1180,
0x0180, 0x1E80, 0x0E80, 0x1680, 0x0680, 0x1A80, 0x0A80, 0x1280,
0x0280, 0x1C80, 0x0C80, 0x1480, 0x0480, 0x1880, 0x0880, 0x1080,
0x0080, 0x1F00, 0x0F00, 0x1700, 0x0700, 0x1B00, 0x0B00, 0x1300,
0x0DA0, 0x05A0, 0x09A0, 0x01A0, 0x0EA0, 0x06A0, 0x0AA0, 0x02A0,
0x0CA0, 0x04A0, 0x08A0, 0x00A0, 0x0F20, 0x0720, 0x0B20, 0x0320,
0x0D20, 0x0520, 0x0920, 0x0120, 0x0E20, 0x0620, 0x0A20, 0x0220,
0x0C20, 0x0420, 0x0820, 0x0020, 0x0FC0, 0x07C0, 0x0BC0, 0x03C0,
0x0DC0, 0x05C0, 0x09C0, 0x01C0, 0x0EC0, 0x06C0, 0x0AC0, 0x02C0,
0x0CC0, 0x04C0, 0x08C0, 0x00C0, 0x0F40, 0x0740, 0x0B40, 0x0340,
0x0300, 0x0D40, 0x1D00, 0x0D00, 0x1500, 0x0540, 0x0500, 0x1900,
0x0900, 0x0940, 0x1100, 0x0100, 0x1E00, 0x0E00, 0x0140, 0x1600,
0x0600, 0x1A00, 0x0E40, 0x0640, 0x0A40, 0x0A00, 0x1200, 0x0200,
0x1C00, 0x0C00, 0x1400, 0x0400, 0x1800, 0x0800, 0x1000, 0x0000
};
//-----------------------------------------------------------------------------
// Local functions
static void GenDecodeTabs(
unsigned char * positions, // [out] Table of positions
unsigned char * start_indexes, // [in] Table of start indexes
unsigned char * length_bits, // [in] Table of lengths. Each length is stored as number of bits
size_t elements) // [in] Number of elements in start_indexes and length_bits
{
unsigned int index;
unsigned int length;
size_t i;
for(i = 0; i < elements; i++)
{
length = 1 << length_bits[i]; // Get the length in bytes
for(index = start_indexes[i]; index < 0x100; index += length)
{
positions[index] = (unsigned char)i;
}
}
}
static void GenAscTabs(TDcmpStruct * pWork)
{
unsigned short * pChCodeAsc = &ChCodeAsc[0xFF];
unsigned int acc, add;
unsigned short count;
for(count = 0x00FF; pChCodeAsc >= ChCodeAsc; pChCodeAsc--, count--)
{
unsigned char * pChBitsAsc = pWork->ChBitsAsc + count;
unsigned char bits_asc = *pChBitsAsc;
if(bits_asc <= 8)
{
add = (1 << bits_asc);
acc = *pChCodeAsc;
do
{
pWork->offs2C34[acc] = (unsigned char)count;
acc += add;
}
while(acc < 0x100);
}
else if((acc = (*pChCodeAsc & 0xFF)) != 0)
{
pWork->offs2C34[acc] = 0xFF;
if(*pChCodeAsc & 0x3F)
{
bits_asc -= 4;
*pChBitsAsc = bits_asc;
add = (1 << bits_asc);
acc = *pChCodeAsc >> 4;
do
{
pWork->offs2D34[acc] = (unsigned char)count;
acc += add;
}
while(acc < 0x100);
}
else
{
bits_asc -= 6;
*pChBitsAsc = bits_asc;
add = (1 << bits_asc);
acc = *pChCodeAsc >> 6;
do
{
pWork->offs2E34[acc] = (unsigned char)count;
acc += add;
}
while(acc < 0x80);
}
}
else
{
bits_asc -= 8;
*pChBitsAsc = bits_asc;
add = (1 << bits_asc);
acc = *pChCodeAsc >> 8;
do
{
pWork->offs2EB4[acc] = (unsigned char)count;
acc += add;
}
while(acc < 0x100);
}
}
}
//-----------------------------------------------------------------------------
// Removes given number of bits in the bit buffer. New bits are reloaded from
// the input buffer, if needed.
// Returns: PKDCL_OK: Operation was successful
// PKDCL_STREAM_END: There are no more bits in the input buffer
static int WasteBits(TDcmpStruct * pWork, unsigned int nBits)
{
// If number of bits required is less than number of (bits in the buffer) ?
if(nBits <= pWork->extra_bits)
{
pWork->extra_bits -= nBits;
pWork->bit_buff >>= nBits;
return PKDCL_OK;
}
// Load input buffer if necessary
pWork->bit_buff >>= pWork->extra_bits;
if(pWork->in_pos == pWork->in_bytes)
{
pWork->in_pos = sizeof(pWork->in_buff);
if((pWork->in_bytes = pWork->read_buf((char *)pWork->in_buff, &pWork->in_pos, pWork->param)) == 0)
return PKDCL_STREAM_END;
pWork->in_pos = 0;
}
// Update bit buffer
pWork->bit_buff |= (pWork->in_buff[pWork->in_pos++] << 8);
pWork->bit_buff >>= (nBits - pWork->extra_bits);
pWork->extra_bits = (pWork->extra_bits - nBits) + 8;
return PKDCL_OK;
}
//-----------------------------------------------------------------------------
// Decodes next literal from the input (compressed) data.
// Returns : 0x000: One byte 0x00
// 0x001: One byte 0x01
// ...
// 0x0FF: One byte 0xFF
// 0x100: Repetition, length of 0x02 bytes
// 0x101: Repetition, length of 0x03 bytes
// ...
// 0x304: Repetition, length of 0x206 bytes
// 0x305: End of stream
// 0x306: Error
static unsigned int DecodeLit(TDcmpStruct * pWork)
{
unsigned int extra_length_bits; // Number of bits of extra literal length
unsigned int length_code; // Length code
unsigned int value;
// Test the current bit in byte buffer. If is not set, simply return the next 8 bits.
if(pWork->bit_buff & 1)
{
// Remove one bit from the input data
if(WasteBits(pWork, 1))
return 0x306;
// The next 8 bits hold the index to the length code table
length_code = pWork->LengthCodes[pWork->bit_buff & 0xFF];
// Remove the apropriate number of bits
if(WasteBits(pWork, pWork->LenBits[length_code]))
return 0x306;
// Are there some extra bits for the obtained length code ?
if((extra_length_bits = pWork->ExLenBits[length_code]) != 0)
{
unsigned int extra_length = pWork->bit_buff & ((1 << extra_length_bits) - 1);
if(WasteBits(pWork, extra_length_bits))
{
if((length_code + extra_length) != 0x10E)
return 0x306;
}
length_code = pWork->LenBase[length_code] + extra_length;
}
// In order to distinguish uncompressed byte from repetition length,
// we have to add 0x100 to the length.
return length_code + 0x100;
}
// Remove one bit from the input data
if(WasteBits(pWork, 1))
return 0x306;
// If the binary compression type, read 8 bits and return them as one byte.
if(pWork->ctype == CMP_BINARY)
{
unsigned int uncompressed_byte = pWork->bit_buff & 0xFF;
if(WasteBits(pWork, 8))
return 0x306;
return uncompressed_byte;
}
// When ASCII compression ...
if(pWork->bit_buff & 0xFF)
{
value = pWork->offs2C34[pWork->bit_buff & 0xFF];
if(value == 0xFF)
{
if(pWork->bit_buff & 0x3F)
{
if(WasteBits(pWork, 4))
return 0x306;
value = pWork->offs2D34[pWork->bit_buff & 0xFF];
}
else
{
if(WasteBits(pWork, 6))
return 0x306;
value = pWork->offs2E34[pWork->bit_buff & 0x7F];
}
}
}
else
{
if(WasteBits(pWork, 8))
return 0x306;
value = pWork->offs2EB4[pWork->bit_buff & 0xFF];
}
return WasteBits(pWork, pWork->ChBitsAsc[value]) ? 0x306 : value;
}
//-----------------------------------------------------------------------------
// Decodes the distance of the repetition, backwards relative to the
// current output buffer position
static unsigned int DecodeDist(TDcmpStruct * pWork, unsigned int rep_length)
{
unsigned int dist_pos_code; // Distance position code
unsigned int dist_pos_bits; // Number of bits of distance position
unsigned int distance; // Distance position
// Next 2-8 bits in the input buffer is the distance position code
dist_pos_code = pWork->DistPosCodes[pWork->bit_buff & 0xFF];
dist_pos_bits = pWork->DistBits[dist_pos_code];
if(WasteBits(pWork, dist_pos_bits))
return 0;
if(rep_length == 2)
{
// If the repetition is only 2 bytes length,
// then take 2 bits from the stream in order to get the distance
distance = (dist_pos_code << 2) | (pWork->bit_buff & 0x03);
if(WasteBits(pWork, 2))
return 0;
}
else
{
// If the repetition is more than 2 bytes length,
// then take "dsize_bits" bits in order to get the distance
distance = (dist_pos_code << pWork->dsize_bits) | (pWork->bit_buff & pWork->dsize_mask);
if(WasteBits(pWork, pWork->dsize_bits))
return 0;
}
return distance + 1;
}
static unsigned int Expand(TDcmpStruct * pWork)
{
unsigned int next_literal; // Literal decoded from the compressed data
unsigned int result; // Value to be returned
unsigned int copyBytes; // Number of bytes to copy to the output buffer
pWork->outputPos = 0x1000; // Initialize output buffer position
// Decode the next literal from the input data.
// The returned literal can either be an uncompressed byte (next_literal < 0x100)
// or an encoded length of the repeating byte sequence that
// is to be copied to the current buffer position
while((result = next_literal = DecodeLit(pWork)) < 0x305)
{
// If the literal is greater than 0x100, it holds length
// of repeating byte sequence
// literal of 0x100 means repeating sequence of 0x2 bytes
// literal of 0x101 means repeating sequence of 0x3 bytes
// ...
// literal of 0x305 means repeating sequence of 0x207 bytes
if(next_literal >= 0x100)
{
unsigned char * source;
unsigned char * target;
unsigned int rep_length; // Length of the repetition, in bytes
unsigned int minus_dist; // Backward distance to the repetition, relative to the current buffer position
// Get the length of the repeating sequence.
// Note that the repeating block may overlap the current output position,
// for example if there was a sequence of equal bytes
rep_length = next_literal - 0xFE;
// Get backward distance to the repetition
if((minus_dist = DecodeDist(pWork, rep_length)) == 0)
{
result = 0x306;
break;
}
// Target and source pointer
target = &pWork->out_buff[pWork->outputPos];
source = target - minus_dist;
// Update buffer output position
pWork->outputPos += rep_length;
// Copy the repeating sequence
while(rep_length-- > 0)
*target++ = *source++;
}
else
{
pWork->out_buff[pWork->outputPos++] = (unsigned char)next_literal;
}
// Flush the output buffer, if number of extracted bytes has reached the end
if(pWork->outputPos >= 0x2000)
{
// Copy decompressed data into user buffer
copyBytes = 0x1000;
pWork->write_buf((char *)&pWork->out_buff[0x1000], &copyBytes, pWork->param);
// Now copy the decompressed data to the first half of the buffer.
// This is needed because the decompression might reuse them as repetitions.
// Note that if the output buffer overflowed previously, the extra decompressed bytes
// are stored in "out_buff_overflow", and they will now be
// within decompressed part of the output buffer.
memmove(pWork->out_buff, &pWork->out_buff[0x1000], pWork->outputPos - 0x1000);
pWork->outputPos -= 0x1000;
}
}
// Flush any remaining decompressed bytes
copyBytes = pWork->outputPos - 0x1000;
pWork->write_buf((char *)&pWork->out_buff[0x1000], &copyBytes, pWork->param);
return result;
}
//-----------------------------------------------------------------------------
// Main exploding function.
unsigned int explode(
unsigned int (*read_buf)(char *buf, unsigned int *size, void *param),
void (*write_buf)(char *buf, unsigned int *size, void *param),
char *work_buf,
void *param)
{
TDcmpStruct * pWork = (TDcmpStruct *)work_buf;
// Initialize work struct and load compressed data
// Note: The caller must zero the "work_buff" before passing it to explode
pWork->read_buf = read_buf;
pWork->write_buf = write_buf;
pWork->param = param;
pWork->in_pos = sizeof(pWork->in_buff);
pWork->in_bytes = pWork->read_buf((char *)pWork->in_buff, &pWork->in_pos, pWork->param);
if(pWork->in_bytes <= 4)
return CMP_BAD_DATA;
pWork->ctype = pWork->in_buff[0]; // Get the compression type (CMP_BINARY or CMP_ASCII)
pWork->dsize_bits = pWork->in_buff[1]; // Get the dictionary size
pWork->bit_buff = pWork->in_buff[2]; // Initialize 16-bit bit buffer
pWork->extra_bits = 0; // Extra (over 8) bits
pWork->in_pos = 3; // Position in input buffer
// Test for the valid dictionary size
if(4 > pWork->dsize_bits || pWork->dsize_bits > 6)
return CMP_INVALID_DICTSIZE;
pWork->dsize_mask = 0xFFFF >> (0x10 - pWork->dsize_bits); // Shifted by 'sar' instruction
if(pWork->ctype != CMP_BINARY)
{
if(pWork->ctype != CMP_ASCII)
return CMP_INVALID_MODE;
memcpy(pWork->ChBitsAsc, ChBitsAsc, sizeof(pWork->ChBitsAsc));
GenAscTabs(pWork);
}
memcpy(pWork->LenBits, LenBits, sizeof(pWork->LenBits));
GenDecodeTabs(pWork->LengthCodes, LenCode, pWork->LenBits, sizeof(pWork->LenBits));
memcpy(pWork->ExLenBits, ExLenBits, sizeof(pWork->ExLenBits));
memcpy(pWork->LenBase, LenBase, sizeof(pWork->LenBase));
memcpy(pWork->DistBits, DistBits, sizeof(pWork->DistBits));
GenDecodeTabs(pWork->DistPosCodes, DistCode, pWork->DistBits, sizeof(pWork->DistBits));
if(Expand(pWork) != 0x306)
return CMP_NO_ERROR;
return CMP_ABORT;
}

775
3rdParty/PKWare/implode.cpp vendored Normal file
View File

@ -0,0 +1,775 @@
/*****************************************************************************/
/* implode.cpp Copyright (c) Ladislav Zezula 2003 */
/*---------------------------------------------------------------------------*/
/* Implode function of PKWARE Data Compression library */
/*---------------------------------------------------------------------------*/
/* Date Ver Who Comment */
/* -------- ---- --- ------- */
/* 11.04.03 1.00 Lad First version of implode.cpp */
/* 02.05.03 1.00 Lad Stress test done */
/* 22.04.10 1.01 Lad Documented */
/*****************************************************************************/
#include <assert.h>
#include <string.h>
#include "pkware.h"
#if ((1200 < _MSC_VER) && (_MSC_VER < 1400))
#pragma optimize("", off)
#endif
//-----------------------------------------------------------------------------
// Defines
#define MAX_REP_LENGTH 0x204 // The longest allowed repetition
static char CopyrightPkware[] = "PKWARE Data Compression Library for Win32\r\n"
"Copyright 1989-1995 PKWARE Inc. All Rights Reserved\r\n"
"Patent No. 5,051,745\r\n"
"PKWARE Data Compression Library Reg. U.S. Pat. and Tm. Off.\r\n"
"Version 1.11\r\n";
//-----------------------------------------------------------------------------
// Tables
static unsigned char DistBits[] =
{
0x02, 0x04, 0x04, 0x05, 0x05, 0x05, 0x05, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06,
0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07, 0x07,
0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08
};
static unsigned char DistCode[] =
{
0x03, 0x0D, 0x05, 0x19, 0x09, 0x11, 0x01, 0x3E, 0x1E, 0x2E, 0x0E, 0x36, 0x16, 0x26, 0x06, 0x3A,
0x1A, 0x2A, 0x0A, 0x32, 0x12, 0x22, 0x42, 0x02, 0x7C, 0x3C, 0x5C, 0x1C, 0x6C, 0x2C, 0x4C, 0x0C,
0x74, 0x34, 0x54, 0x14, 0x64, 0x24, 0x44, 0x04, 0x78, 0x38, 0x58, 0x18, 0x68, 0x28, 0x48, 0x08,
0xF0, 0x70, 0xB0, 0x30, 0xD0, 0x50, 0x90, 0x10, 0xE0, 0x60, 0xA0, 0x20, 0xC0, 0x40, 0x80, 0x00
};
static unsigned char ExLenBits[] =
{
0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08
};
static unsigned char LenBits[] =
{
0x03, 0x02, 0x03, 0x03, 0x04, 0x04, 0x04, 0x05, 0x05, 0x05, 0x05, 0x06, 0x06, 0x06, 0x07, 0x07
};
static unsigned char LenCode[] =
{
0x05, 0x03, 0x01, 0x06, 0x0A, 0x02, 0x0C, 0x14, 0x04, 0x18, 0x08, 0x30, 0x10, 0x20, 0x40, 0x00
};
static unsigned char ChBitsAsc[] =
{
0x0B, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x08, 0x07, 0x0C, 0x0C, 0x07, 0x0C, 0x0C,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0D, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x04, 0x0A, 0x08, 0x0C, 0x0A, 0x0C, 0x0A, 0x08, 0x07, 0x07, 0x08, 0x09, 0x07, 0x06, 0x07, 0x08,
0x07, 0x06, 0x07, 0x07, 0x07, 0x07, 0x08, 0x07, 0x07, 0x08, 0x08, 0x0C, 0x0B, 0x07, 0x09, 0x0B,
0x0C, 0x06, 0x07, 0x06, 0x06, 0x05, 0x07, 0x08, 0x08, 0x06, 0x0B, 0x09, 0x06, 0x07, 0x06, 0x06,
0x07, 0x0B, 0x06, 0x06, 0x06, 0x07, 0x09, 0x08, 0x09, 0x09, 0x0B, 0x08, 0x0B, 0x09, 0x0C, 0x08,
0x0C, 0x05, 0x06, 0x06, 0x06, 0x05, 0x06, 0x06, 0x06, 0x05, 0x0B, 0x07, 0x05, 0x06, 0x05, 0x05,
0x06, 0x0A, 0x05, 0x05, 0x05, 0x05, 0x08, 0x07, 0x08, 0x08, 0x0A, 0x0B, 0x0B, 0x0C, 0x0C, 0x0C,
0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D,
0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D,
0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C,
0x0D, 0x0C, 0x0D, 0x0D, 0x0D, 0x0C, 0x0D, 0x0D, 0x0D, 0x0C, 0x0D, 0x0D, 0x0D, 0x0D, 0x0C, 0x0D,
0x0D, 0x0D, 0x0C, 0x0C, 0x0C, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D, 0x0D
};
static unsigned short ChCodeAsc[] =
{
0x0490, 0x0FE0, 0x07E0, 0x0BE0, 0x03E0, 0x0DE0, 0x05E0, 0x09E0,
0x01E0, 0x00B8, 0x0062, 0x0EE0, 0x06E0, 0x0022, 0x0AE0, 0x02E0,
0x0CE0, 0x04E0, 0x08E0, 0x00E0, 0x0F60, 0x0760, 0x0B60, 0x0360,
0x0D60, 0x0560, 0x1240, 0x0960, 0x0160, 0x0E60, 0x0660, 0x0A60,
0x000F, 0x0250, 0x0038, 0x0260, 0x0050, 0x0C60, 0x0390, 0x00D8,
0x0042, 0x0002, 0x0058, 0x01B0, 0x007C, 0x0029, 0x003C, 0x0098,
0x005C, 0x0009, 0x001C, 0x006C, 0x002C, 0x004C, 0x0018, 0x000C,
0x0074, 0x00E8, 0x0068, 0x0460, 0x0090, 0x0034, 0x00B0, 0x0710,
0x0860, 0x0031, 0x0054, 0x0011, 0x0021, 0x0017, 0x0014, 0x00A8,
0x0028, 0x0001, 0x0310, 0x0130, 0x003E, 0x0064, 0x001E, 0x002E,
0x0024, 0x0510, 0x000E, 0x0036, 0x0016, 0x0044, 0x0030, 0x00C8,
0x01D0, 0x00D0, 0x0110, 0x0048, 0x0610, 0x0150, 0x0060, 0x0088,
0x0FA0, 0x0007, 0x0026, 0x0006, 0x003A, 0x001B, 0x001A, 0x002A,
0x000A, 0x000B, 0x0210, 0x0004, 0x0013, 0x0032, 0x0003, 0x001D,
0x0012, 0x0190, 0x000D, 0x0015, 0x0005, 0x0019, 0x0008, 0x0078,
0x00F0, 0x0070, 0x0290, 0x0410, 0x0010, 0x07A0, 0x0BA0, 0x03A0,
0x0240, 0x1C40, 0x0C40, 0x1440, 0x0440, 0x1840, 0x0840, 0x1040,
0x0040, 0x1F80, 0x0F80, 0x1780, 0x0780, 0x1B80, 0x0B80, 0x1380,
0x0380, 0x1D80, 0x0D80, 0x1580, 0x0580, 0x1980, 0x0980, 0x1180,
0x0180, 0x1E80, 0x0E80, 0x1680, 0x0680, 0x1A80, 0x0A80, 0x1280,
0x0280, 0x1C80, 0x0C80, 0x1480, 0x0480, 0x1880, 0x0880, 0x1080,
0x0080, 0x1F00, 0x0F00, 0x1700, 0x0700, 0x1B00, 0x0B00, 0x1300,
0x0DA0, 0x05A0, 0x09A0, 0x01A0, 0x0EA0, 0x06A0, 0x0AA0, 0x02A0,
0x0CA0, 0x04A0, 0x08A0, 0x00A0, 0x0F20, 0x0720, 0x0B20, 0x0320,
0x0D20, 0x0520, 0x0920, 0x0120, 0x0E20, 0x0620, 0x0A20, 0x0220,
0x0C20, 0x0420, 0x0820, 0x0020, 0x0FC0, 0x07C0, 0x0BC0, 0x03C0,
0x0DC0, 0x05C0, 0x09C0, 0x01C0, 0x0EC0, 0x06C0, 0x0AC0, 0x02C0,
0x0CC0, 0x04C0, 0x08C0, 0x00C0, 0x0F40, 0x0740, 0x0B40, 0x0340,
0x0300, 0x0D40, 0x1D00, 0x0D00, 0x1500, 0x0540, 0x0500, 0x1900,
0x0900, 0x0940, 0x1100, 0x0100, 0x1E00, 0x0E00, 0x0140, 0x1600,
0x0600, 0x1A00, 0x0E40, 0x0640, 0x0A40, 0x0A00, 0x1200, 0x0200,
0x1C00, 0x0C00, 0x1400, 0x0400, 0x1800, 0x0800, 0x1000, 0x0000
};
//-----------------------------------------------------------------------------
// Macros
// Macro for calculating hash of the current byte pair.
// Note that most exact byte pair hash would be buffer[0] + buffer[1] << 0x08,
// but even this way gives nice indication of equal byte pairs, with significantly
// smaller size of the array that holds numbers of those hashes
#define BYTE_PAIR_HASH(buffer) ((buffer[0] * 4) + (buffer[1] * 5))
//-----------------------------------------------------------------------------
// Local functions
// Builds the "hash_to_index" table and "pair_hash_offsets" table.
// Every element of "hash_to_index" will contain lowest index to the
// "pair_hash_offsets" table, effectively giving offset of the first
// occurence of the given PAIR_HASH in the input data.
static void SortBuffer(TCmpStruct * pWork, unsigned char * buffer_begin, unsigned char * buffer_end)
{
unsigned short * phash_to_index;
unsigned char * buffer_ptr;
unsigned short total_sum = 0;
unsigned long byte_pair_hash; // Hash value of the byte pair
unsigned short byte_pair_offs; // Offset of the byte pair, relative to "work_buff"
// Zero the entire "phash_to_index" table
memset(pWork->phash_to_index, 0, sizeof(pWork->phash_to_index));
// Step 1: Count amount of each PAIR_HASH in the input buffer
// The table will look like this:
// offs 0x000: Number of occurences of PAIR_HASH 0
// offs 0x001: Number of occurences of PAIR_HASH 1
// ...
// offs 0x8F7: Number of occurences of PAIR_HASH 0x8F7 (the highest hash value)
for(buffer_ptr = buffer_begin; buffer_ptr < buffer_end; buffer_ptr++)
pWork->phash_to_index[BYTE_PAIR_HASH(buffer_ptr)]++;
// Step 2: Convert the table to the array of PAIR_HASH amounts.
// Each element contains count of PAIR_HASHes that is less or equal
// to element index
// The table will look like this:
// offs 0x000: Number of occurences of PAIR_HASH 0 or lower
// offs 0x001: Number of occurences of PAIR_HASH 1 or lower
// ...
// offs 0x8F7: Number of occurences of PAIR_HASH 0x8F7 or lower
for(phash_to_index = pWork->phash_to_index; phash_to_index < &pWork->phash_to_index_end; phash_to_index++)
{
total_sum = total_sum + phash_to_index[0];
phash_to_index[0] = total_sum;
}
// Step 3: Convert the table to the array of indexes.
// Now, each element contains index to the first occurence of given PAIR_HASH
for(buffer_end--; buffer_end >= buffer_begin; buffer_end--)
{
byte_pair_hash = BYTE_PAIR_HASH(buffer_end);
byte_pair_offs = (unsigned short)(buffer_end - pWork->work_buff);
pWork->phash_to_index[byte_pair_hash]--;
pWork->phash_offs[pWork->phash_to_index[byte_pair_hash]] = byte_pair_offs;
}
}
static void FlushBuf(TCmpStruct * pWork)
{
unsigned char save_ch1;
unsigned char save_ch2;
unsigned int size = 0x800;
pWork->write_buf(pWork->out_buff, &size, pWork->param);
save_ch1 = pWork->out_buff[0x800];
save_ch2 = pWork->out_buff[pWork->out_bytes];
pWork->out_bytes -= 0x800;
memset(pWork->out_buff, 0, sizeof(pWork->out_buff));
if(pWork->out_bytes != 0)
pWork->out_buff[0] = save_ch1;
if(pWork->out_bits != 0)
pWork->out_buff[pWork->out_bytes] = save_ch2;
}
static void OutputBits(TCmpStruct * pWork, unsigned int nbits, unsigned long bit_buff)
{
unsigned int out_bits;
// If more than 8 bits to output, do recursion
if(nbits > 8)
{
OutputBits(pWork, 8, bit_buff);
bit_buff >>= 8;
nbits -= 8;
}
// Add bits to the last out byte in out_buff;
out_bits = pWork->out_bits;
pWork->out_buff[pWork->out_bytes] |= (unsigned char)(bit_buff << out_bits);
pWork->out_bits += nbits;
// If 8 or more bits, increment number of bytes
if(pWork->out_bits > 8)
{
pWork->out_bytes++;
bit_buff >>= (8 - out_bits);
pWork->out_buff[pWork->out_bytes] = (unsigned char)bit_buff;
pWork->out_bits &= 7;
}
else
{
pWork->out_bits &= 7;
if(pWork->out_bits == 0)
pWork->out_bytes++;
}
// If there is enough compressed bytes, flush them
if(pWork->out_bytes >= 0x800)
FlushBuf(pWork);
}
// This function searches for a repetition
// (a previous occurence of the current byte sequence)
// Returns length of the repetition, and stores the backward distance
// to pWork structure.
static unsigned int FindRep(TCmpStruct * pWork, unsigned char * input_data)
{
unsigned short * phash_to_index; // Pointer into pWork->phash_to_index table
unsigned short * phash_offs; // Pointer to the table containing offsets of each PAIR_HASH
unsigned char * repetition_limit; // An eventual repetition must be at position below this pointer
unsigned char * prev_repetition; // Pointer to the previous occurence of the current PAIR_HASH
unsigned char * prev_rep_end; // End of the previous repetition
unsigned char * input_data_ptr;
unsigned short phash_offs_index; // Index to the table with PAIR_HASH positions
unsigned short min_phash_offs; // The lowest allowed hash offset
unsigned short offs_in_rep; // Offset within found repetition
unsigned int equal_byte_count; // Number of bytes that are equal to the previous occurence
unsigned int rep_length = 1; // Length of the found repetition
unsigned int rep_length2; // Secondary repetition
unsigned char pre_last_byte; // Last but one byte from a repetion
unsigned short di_val;
// Calculate the previous position of the PAIR_HASH
phash_to_index = pWork->phash_to_index + BYTE_PAIR_HASH(input_data);
min_phash_offs = (unsigned short)((input_data - pWork->work_buff) - pWork->dsize_bytes + 1);
phash_offs_index = phash_to_index[0];
// If the PAIR_HASH offset is below the limit, find a next one
phash_offs = pWork->phash_offs + phash_offs_index;
if(*phash_offs < min_phash_offs)
{
while(*phash_offs < min_phash_offs)
{
phash_offs_index++;
phash_offs++;
}
*phash_to_index = phash_offs_index;
}
// Get the first location of the PAIR_HASH,
// and thus the first eventual location of byte repetition
phash_offs = pWork->phash_offs + phash_offs_index;
prev_repetition = pWork->work_buff + phash_offs[0];
repetition_limit = input_data - 1;
// If the current PAIR_HASH was not encountered before,
// we haven't found a repetition.
if(prev_repetition >= repetition_limit)
return 0;
// We have found a match of a PAIR_HASH. Now we have to make sure
// that it is also a byte match, because PAIR_HASH is not unique.
// We compare the bytes and count the length of the repetition
input_data_ptr = input_data;
for(;;)
{
// If the first byte of the repetition and the so-far-last byte
// of the repetition are equal, we will compare the blocks.
if(*input_data_ptr == *prev_repetition && input_data_ptr[rep_length-1] == prev_repetition[rep_length-1])
{
// Skip the current byte
prev_repetition++;
input_data_ptr++;
equal_byte_count = 2;
// Now count how many more bytes are equal
while(equal_byte_count < MAX_REP_LENGTH)
{
prev_repetition++;
input_data_ptr++;
// Are the bytes different ?
if(*prev_repetition != *input_data_ptr)
break;
equal_byte_count++;
}
// If we found a repetition of at least the same length, take it.
// If there are multiple repetitions in the input buffer, this will
// make sure that we find the most recent one, which in turn allows
// us to store backward length in less amount of bits
input_data_ptr = input_data;
if(equal_byte_count >= rep_length)
{
// Calculate the backward distance of the repetition.
// Note that the distance is stored as decremented by 1
pWork->distance = (unsigned int)(input_data - prev_repetition + equal_byte_count - 1);
// Repetitions longer than 10 bytes will be stored in more bits,
// so they need a bit different handling
if((rep_length = equal_byte_count) > 10)
break;
}
}
// Move forward in the table of PAIR_HASH repetitions.
// There might be a more recent occurence of the same repetition.
phash_offs_index++;
phash_offs++;
prev_repetition = pWork->work_buff + phash_offs[0];
// If the next repetition is beyond the minimum allowed repetition, we are done.
if(prev_repetition >= repetition_limit)
{
// A repetition must have at least 2 bytes, otherwise it's not worth it
return (rep_length >= 2) ? rep_length : 0;
}
}
// If the repetition has max length of 0x204 bytes, we can't go any fuhrter
if(equal_byte_count == MAX_REP_LENGTH)
{
pWork->distance--;
return equal_byte_count;
}
// Check for possibility of a repetition that occurs at more recent position
phash_offs = pWork->phash_offs + phash_offs_index;
if(pWork->work_buff + phash_offs[1] >= repetition_limit)
return rep_length;
//
// The following part checks if there isn't a longer repetition at
// a latter offset, that would lead to better compression.
//
// Example of data that can trigger this optimization:
//
// "EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEQQQQQQQQQQQQ"
// "XYZ"
// "EEEEEEEEEEEEEEEEQQQQQQQQQQQQ";
//
// Description of data in this buffer
// [0x00] Single byte "E"
// [0x01] Single byte "E"
// [0x02] Repeat 0x1E bytes from [0x00]
// [0x20] Single byte "X"
// [0x21] Single byte "Y"
// [0x22] Single byte "Z"
// [0x23] 17 possible previous repetitions of length at least 0x10 bytes:
// - Repetition of 0x10 bytes from [0x00] "EEEEEEEEEEEEEEEE"
// - Repetition of 0x10 bytes from [0x01] "EEEEEEEEEEEEEEEE"
// - Repetition of 0x10 bytes from [0x02] "EEEEEEEEEEEEEEEE"
// ...
// - Repetition of 0x10 bytes from [0x0F] "EEEEEEEEEEEEEEEE"
// - Repetition of 0x1C bytes from [0x10] "EEEEEEEEEEEEEEEEQQQQQQQQQQQQ"
// The last repetition is the best one.
//
pWork->offs09BC[0] = 0xFFFF;
pWork->offs09BC[1] = 0x0000;
di_val = 0;
// Note: I failed to figure out what does the table "offs09BC" mean.
// If anyone has an idea, let me know to zezula_at_volny_dot_cz
for(offs_in_rep = 1; offs_in_rep < rep_length; )
{
if(input_data[offs_in_rep] != input_data[di_val])
{
di_val = pWork->offs09BC[di_val];
if(di_val != 0xFFFF)
continue;
}
pWork->offs09BC[++offs_in_rep] = ++di_val;
}
//
// Now go through all the repetitions from the first found one
// to the current input data, and check if any of them migh be
// a start of a greater sequence match.
//
prev_repetition = pWork->work_buff + phash_offs[0];
prev_rep_end = prev_repetition + rep_length;
rep_length2 = rep_length;
for(;;)
{
rep_length2 = pWork->offs09BC[rep_length2];
if(rep_length2 == 0xFFFF)
rep_length2 = 0;
// Get the pointer to the previous repetition
phash_offs = pWork->phash_offs + phash_offs_index;
// Skip those repetitions that don't reach the end
// of the first found repetition
do
{
phash_offs++;
phash_offs_index++;
prev_repetition = pWork->work_buff + *phash_offs;
if(prev_repetition >= repetition_limit)
return rep_length;
}
while(prev_repetition + rep_length2 < prev_rep_end);
// Verify if the last but one byte from the repetition matches
// the last but one byte from the input data.
// If not, find a next repetition
pre_last_byte = input_data[rep_length - 2];
if(pre_last_byte == prev_repetition[rep_length - 2])
{
// If the new repetition reaches beyond the end
// of previously found repetition, reset the repetition length to zero.
if(prev_repetition + rep_length2 != prev_rep_end)
{
prev_rep_end = prev_repetition;
rep_length2 = 0;
}
}
else
{
phash_offs = pWork->phash_offs + phash_offs_index;
do
{
phash_offs++;
phash_offs_index++;
prev_repetition = pWork->work_buff + *phash_offs;
if(prev_repetition >= repetition_limit)
return rep_length;
}
while(prev_repetition[rep_length - 2] != pre_last_byte || prev_repetition[0] != input_data[0]);
// Reset the length of the repetition to 2 bytes only
prev_rep_end = prev_repetition + 2;
rep_length2 = 2;
}
// Find out how many more characters are equal to the first repetition.
while(*prev_rep_end == input_data[rep_length2])
{
if(++rep_length2 >= 0x204)
break;
prev_rep_end++;
}
// Is the newly found repetion at least as long as the previous one ?
if(rep_length2 >= rep_length)
{
// Calculate the distance of the new repetition
pWork->distance = (unsigned int)(input_data - prev_repetition - 1);
if((rep_length = rep_length2) == 0x204)
return rep_length;
// Update the additional elements in the "offs09BC" table
// to reflect new rep length
while(offs_in_rep < rep_length2)
{
if(input_data[offs_in_rep] != input_data[di_val])
{
di_val = pWork->offs09BC[di_val];
if(di_val != 0xFFFF)
continue;
}
pWork->offs09BC[++offs_in_rep] = ++di_val;
}
}
}
}
static void WriteCmpData(TCmpStruct * pWork)
{
unsigned char * input_data_end; // Pointer to the end of the input data
unsigned char * input_data = pWork->work_buff + pWork->dsize_bytes + 0x204;
unsigned int input_data_ended = 0; // If 1, then all data from the input stream have been already loaded
unsigned int save_rep_length; // Saved length of current repetition
unsigned int save_distance = 0; // Saved distance of current repetition
unsigned int rep_length; // Length of the found repetition
unsigned int phase = 0; //
// Store the compression type and dictionary size
pWork->out_buff[0] = (char)pWork->ctype;
pWork->out_buff[1] = (char)pWork->dsize_bits;
pWork->out_bytes = 2;
// Reset output buffer to zero
memset(&pWork->out_buff[2], 0, sizeof(pWork->out_buff) - 2);
pWork->out_bits = 0;
while(input_data_ended == 0)
{
unsigned int bytes_to_load = 0x1000;
int total_loaded = 0;
int bytes_loaded;
// Load the bytes from the input stream, up to 0x1000 bytes
while(bytes_to_load != 0)
{
bytes_loaded = pWork->read_buf((char *)pWork->work_buff + pWork->dsize_bytes + 0x204 + total_loaded,
&bytes_to_load,
pWork->param);
if(bytes_loaded == 0)
{
if(total_loaded == 0 && phase == 0)
goto __Exit;
input_data_ended = 1;
break;
}
else
{
bytes_to_load -= bytes_loaded;
total_loaded += bytes_loaded;
}
}
input_data_end = pWork->work_buff + pWork->dsize_bytes + total_loaded;
if(input_data_ended)
input_data_end += 0x204;
//
// Warning: The end of the buffer passed to "SortBuffer" is actually 2 bytes beyond
// valid data. It is questionable if this is actually a bug or not,
// but it might cause the compressed data output to be dependent on random bytes
// that are in the buffer.
// To prevent that, the calling application must always zero the compression
// buffer before passing it to "implode"
//
// Search the PAIR_HASHes of the loaded blocks. Also, include
// previously compressed data, if any.
switch(phase)
{
case 0:
SortBuffer(pWork, input_data, input_data_end + 1);
phase++;
if(pWork->dsize_bytes != 0x1000)
phase++;
break;
case 1:
SortBuffer(pWork, input_data - pWork->dsize_bytes + 0x204, input_data_end + 1);
phase++;
break;
default:
SortBuffer(pWork, input_data - pWork->dsize_bytes, input_data_end + 1);
break;
}
// Perform the compression of the current block
while(input_data < input_data_end)
{
// Find if the current byte sequence wasn't there before.
rep_length = FindRep(pWork, input_data);
while(rep_length != 0)
{
// If we found repetition of 2 bytes, that is 0x100 or fuhrter back,
// don't bother. Storing the distance of 0x100 bytes would actually
// take more space than storing the 2 bytes as-is.
if(rep_length == 2 && pWork->distance >= 0x100)
break;
// When we are at the end of the input data, we cannot allow
// the repetition to go past the end of the input data.
if(input_data_ended && input_data + rep_length > input_data_end)
{
// Shorten the repetition length so that it only covers valid data
rep_length = (unsigned long)(input_data_end - input_data);
if(rep_length < 2)
break;
// If we got repetition of 2 bytes, that is 0x100 or more backward, don't bother
if(rep_length == 2 && pWork->distance >= 0x100)
break;
goto __FlushRepetition;
}
if(rep_length >= 8 || input_data + 1 >= input_data_end)
goto __FlushRepetition;
// Try to find better repetition 1 byte later.
// Example: "ARROCKFORT" "AROCKFORT"
// When "input_data" points to the second string, FindRep
// returns the occurence of "AR". But there is longer repetition "ROCKFORT",
// beginning 1 byte after.
save_rep_length = rep_length;
save_distance = pWork->distance;
rep_length = FindRep(pWork, input_data + 1);
// Only use the new repetition if it's length is greater than the previous one
if(rep_length > save_rep_length)
{
// If the new repetition if only 1 byte better
// and the previous distance is less than 0x80 bytes, use the previous repetition
if(rep_length > save_rep_length + 1 || save_distance > 0x80)
{
// Flush one byte, so that input_data will point to the secondary repetition
OutputBits(pWork, pWork->nChBits[*input_data], pWork->nChCodes[*input_data]);
input_data++;
continue;
}
}
// Revert to the previous repetition
rep_length = save_rep_length;
pWork->distance = save_distance;
__FlushRepetition:
OutputBits(pWork, pWork->nChBits[rep_length + 0xFE], pWork->nChCodes[rep_length + 0xFE]);
if(rep_length == 2)
{
OutputBits(pWork, pWork->dist_bits[pWork->distance >> 2],
pWork->dist_codes[pWork->distance >> 2]);
OutputBits(pWork, 2, pWork->distance & 3);
}
else
{
OutputBits(pWork, pWork->dist_bits[pWork->distance >> pWork->dsize_bits],
pWork->dist_codes[pWork->distance >> pWork->dsize_bits]);
OutputBits(pWork, pWork->dsize_bits, pWork->dsize_mask & pWork->distance);
}
// Move the begin of the input data by the length of the repetition
input_data += rep_length;
goto _00402252;
}
// If there was no previous repetition for the current position in the input data,
// just output the 9-bit literal for the one character
OutputBits(pWork, pWork->nChBits[*input_data], pWork->nChCodes[*input_data]);
input_data++;
_00402252:;
}
if(input_data_ended == 0)
{
input_data -= 0x1000;
memmove(pWork->work_buff, pWork->work_buff + 0x1000, pWork->dsize_bytes + 0x204);
}
}
__Exit:
// Write the termination literal
OutputBits(pWork, pWork->nChBits[0x305], pWork->nChCodes[0x305]);
if(pWork->out_bits != 0)
pWork->out_bytes++;
pWork->write_buf(pWork->out_buff, &pWork->out_bytes, pWork->param);
return;
}
//-----------------------------------------------------------------------------
// Main imploding function
unsigned int PKEXPORT implode(
unsigned int (*read_buf)(char *buf, unsigned int *size, void *param),
void (*write_buf)(char *buf, unsigned int *size, void *param),
char *work_buf,
void *param,
unsigned int *type,
unsigned int *dsize)
{
TCmpStruct * pWork = (TCmpStruct *)work_buf;
unsigned int nChCode;
unsigned int nCount;
unsigned int i;
int nCount2;
// Fill the work buffer information
// Note: The caller must zero the "work_buff" before passing it to implode
pWork->read_buf = read_buf;
pWork->write_buf = write_buf;
pWork->dsize_bytes = *dsize;
pWork->ctype = *type;
pWork->param = param;
pWork->dsize_bits = 4;
pWork->dsize_mask = 0x0F;
// Test dictionary size
switch(*dsize)
{
case CMP_IMPLODE_DICT_SIZE3: // 0x1000 bytes
pWork->dsize_bits++;
pWork->dsize_mask |= 0x20;
// No break here !!!
case CMP_IMPLODE_DICT_SIZE2: // 0x800 bytes
pWork->dsize_bits++;
pWork->dsize_mask |= 0x10;
// No break here !!!
case CMP_IMPLODE_DICT_SIZE1: // 0x400
break;
default:
return CMP_INVALID_DICTSIZE;
}
// Test the compression type
switch(*type)
{
case CMP_BINARY: // We will compress data with binary compression type
for(nChCode = 0, nCount = 0; nCount < 0x100; nCount++)
{
pWork->nChBits[nCount] = 9;
pWork->nChCodes[nCount] = (unsigned short)nChCode;
nChCode = (nChCode & 0x0000FFFF) + 2;
}
break;
case CMP_ASCII: // We will compress data with ASCII compression type
for(nCount = 0; nCount < 0x100; nCount++)
{
pWork->nChBits[nCount] = (unsigned char )(ChBitsAsc[nCount] + 1);
pWork->nChCodes[nCount] = (unsigned short)(ChCodeAsc[nCount] * 2);
}
break;
default:
return CMP_INVALID_MODE;
}
for(i = 0; i < 0x10; i++)
{
if(1 << ExLenBits[i])
{
for(nCount2 = 0; nCount2 < (1 << ExLenBits[i]); nCount2++)
{
pWork->nChBits[nCount] = (unsigned char)(ExLenBits[i] + LenBits[i] + 1);
pWork->nChCodes[nCount] = (unsigned short)((nCount2 << (LenBits[i] + 1)) | ((LenCode[i] & 0xFFFF00FF) * 2) | 1);
nCount++;
}
}
}
// Copy the distance codes and distance bits and perform the compression
memcpy(&pWork->dist_codes, DistCode, sizeof(DistCode));
memcpy(&pWork->dist_bits, DistBits, sizeof(DistBits));
WriteCmpData(pWork);
return CMP_NO_ERROR;
}

142
3rdParty/PKWare/pkware.h vendored Normal file
View File

@ -0,0 +1,142 @@
/*****************************************************************************/
/* pkware.h Copyright (c) Ladislav Zezula 2003 */
/*---------------------------------------------------------------------------*/
/* Header file for PKWARE Data Compression Library */
/*---------------------------------------------------------------------------*/
/* Date Ver Who Comment */
/* -------- ---- --- ------- */
/* 31.03.03 1.00 Lad The first version of pkware.h */
/*****************************************************************************/
#ifndef __PKWARE_H__
#define __PKWARE_H__
//-----------------------------------------------------------------------------
// Defines
#define CMP_BINARY 0 // Binary compression
#define CMP_ASCII 1 // Ascii compression
#define CMP_NO_ERROR 0
#define CMP_INVALID_DICTSIZE 1
#define CMP_INVALID_MODE 2
#define CMP_BAD_DATA 3
#define CMP_ABORT 4
#define CMP_IMPLODE_DICT_SIZE1 1024 // Dictionary size of 1024
#define CMP_IMPLODE_DICT_SIZE2 2048 // Dictionary size of 2048
#define CMP_IMPLODE_DICT_SIZE3 4096 // Dictionary size of 4096
//-----------------------------------------------------------------------------
// Define calling convention
#ifndef PKEXPORT
#ifdef WIN32
#define PKEXPORT __cdecl // Use for normal __cdecl calling
#else
#define PKEXPORT
#endif
#endif
//-----------------------------------------------------------------------------
// Internal structures
// Compression structure
typedef struct
{
unsigned int distance; // 0000: Backward distance of the currently found repetition, decreased by 1
unsigned int out_bytes; // 0004: # bytes available in out_buff
unsigned int out_bits; // 0008: # of bits available in the last out byte
unsigned int dsize_bits; // 000C: Number of bits needed for dictionary size. 4 = 0x400, 5 = 0x800, 6 = 0x1000
unsigned int dsize_mask; // 0010: Bit mask for dictionary. 0x0F = 0x400, 0x1F = 0x800, 0x3F = 0x1000
unsigned int ctype; // 0014: Compression type (CMP_ASCII or CMP_BINARY)
unsigned int dsize_bytes; // 0018: Dictionary size in bytes
unsigned char dist_bits[0x40]; // 001C: Distance bits
unsigned char dist_codes[0x40]; // 005C: Distance codes
unsigned char nChBits[0x306]; // 009C: Table of literal bit lengths to be put to the output stream
unsigned short nChCodes[0x306]; // 03A2: Table of literal codes to be put to the output stream
unsigned short offs09AE; // 09AE:
void * param; // 09B0: User parameter
unsigned int (*read_buf)(char *buf, unsigned int *size, void *param); // 9B4
void (*write_buf)(char *buf, unsigned int *size, void *param); // 9B8
unsigned short offs09BC[0x204]; // 09BC:
unsigned long offs0DC4; // 0DC4:
unsigned short phash_to_index[0x900]; // 0DC8: Array of indexes (one for each PAIR_HASH) to the "pair_hash_offsets" table
unsigned short phash_to_index_end; // 1FC8: End marker for "phash_to_index" table
char out_buff[0x802]; // 1FCA: Compressed data
unsigned char work_buff[0x2204]; // 27CC: Work buffer
// + DICT_OFFSET => Dictionary
// + UNCMP_OFFSET => Uncompressed data
unsigned short phash_offs[0x2204]; // 49D0: Table of offsets for each PAIR_HASH
} TCmpStruct;
#define CMP_BUFFER_SIZE sizeof(TCmpStruct) // Size of compression structure.
// Defined as 36312 in pkware header file
// Decompression structure
typedef struct
{
unsigned long offs0000; // 0000
unsigned long ctype; // 0004: Compression type (CMP_BINARY or CMP_ASCII)
unsigned long outputPos; // 0008: Position in output buffer
unsigned long dsize_bits; // 000C: Dict size (4, 5, 6 for 0x400, 0x800, 0x1000)
unsigned long dsize_mask; // 0010: Dict size bitmask (0x0F, 0x1F, 0x3F for 0x400, 0x800, 0x1000)
unsigned long bit_buff; // 0014: 16-bit buffer for processing input data
unsigned long extra_bits; // 0018: Number of extra (above 8) bits in bit buffer
unsigned int in_pos; // 001C: Position in in_buff
unsigned long in_bytes; // 0020: Number of bytes in input buffer
void * param; // 0024: Custom parameter
unsigned int (*read_buf)(char *buf, unsigned int *size, void *param); // Pointer to function that reads data from the input stream
void (*write_buf)(char *buf, unsigned int *size, void *param);// Pointer to function that writes data to the output stream
unsigned char out_buff[0x2204]; // 0030: Output circle buffer.
// 0x0000 - 0x0FFF: Previous uncompressed data, kept for repetitions
// 0x1000 - 0x1FFF: Currently decompressed data
// 0x2000 - 0x2203: Reserve space for the longest possible repetition
unsigned char in_buff[0x800]; // 2234: Buffer for data to be decompressed
unsigned char DistPosCodes[0x100]; // 2A34: Table of distance position codes
unsigned char LengthCodes[0x100]; // 2B34: Table of length codes
unsigned char offs2C34[0x100]; // 2C34: Buffer for
unsigned char offs2D34[0x100]; // 2D34: Buffer for
unsigned char offs2E34[0x80]; // 2EB4: Buffer for
unsigned char offs2EB4[0x100]; // 2EB4: Buffer for
unsigned char ChBitsAsc[0x100]; // 2FB4: Buffer for
unsigned char DistBits[0x40]; // 30B4: Numbers of bytes to skip copied block length
unsigned char LenBits[0x10]; // 30F4: Numbers of bits for skip copied block length
unsigned char ExLenBits[0x10]; // 3104: Number of valid bits for copied block
unsigned short LenBase[0x10]; // 3114: Buffer for
} TDcmpStruct;
#define EXP_BUFFER_SIZE sizeof(TDcmpStruct) // Size of decompression structure
// Defined as 12596 in pkware headers
//-----------------------------------------------------------------------------
// Public functions
#ifdef __cplusplus
extern "C" {
#endif
unsigned int PKEXPORT implode(
unsigned int (*read_buf)(char *buf, unsigned int *size, void *param),
void (*write_buf)(char *buf, unsigned int *size, void *param),
char *work_buf,
void *param,
unsigned int *type,
unsigned int *dsize);
unsigned int PKEXPORT explode(
unsigned int (*read_buf)(char *buf, unsigned int *size, void *param),
void (*write_buf)(char *buf, unsigned int *size, void *param),
char *work_buf,
void *param);
#ifdef __cplusplus
} // End of 'extern "C"' declaration
#endif
#endif // __PKWARE_H__