diff options
| author | briaguya <70942617+briaguya-ai@users.noreply.github.com> | 2023-10-17 22:04:12 -0400 |
|---|---|---|
| committer | briaguya <70942617+briaguya-ai@users.noreply.github.com> | 2023-10-17 22:04:12 -0400 |
| commit | e9f4fc638ed904cc0dcdac24ea7708c9a220a702 (patch) | |
| tree | 114c3d432697e04831741e623ae2340f6239bbe9 /ZAPD/yaz0 | |
| parent | dcf56d93ecbd4db3bfd8a4832d7f991467964b98 (diff) | |
update to latest version used by SoH
Diffstat (limited to 'ZAPD/yaz0')
| -rw-r--r-- | ZAPD/yaz0/readwrite.h | 29 | ||||
| -rw-r--r-- | ZAPD/yaz0/yaz0.cpp | 227 | ||||
| -rw-r--r-- | ZAPD/yaz0/yaz0.h | 6 |
3 files changed, 262 insertions, 0 deletions
diff --git a/ZAPD/yaz0/readwrite.h b/ZAPD/yaz0/readwrite.h new file mode 100644 index 0000000..af34715 --- /dev/null +++ b/ZAPD/yaz0/readwrite.h @@ -0,0 +1,29 @@ +#ifndef __READWRITE_H__ +#define __READWRITE_H__ + +#include <stdint.h> + +/* variables */ +union { + uint32_t u; + float f; +} __u32_f32_union__; + +#define U32(x) \ + ((uint32_t)((((uint8_t*)(x))[0] << 24) | (((uint8_t*)(x))[1] << 16) | \ + (((uint8_t*)(x))[2] << 8) | ((uint8_t*)(x))[3])) +#define U16(x) ((uint16_t)(((*((uint8_t*)(x))) << 8) | ((uint8_t*)(x))[1])) +#define U8(x) ((uint8_t)((uint8_t*)(x))[0]) +#define S32(x) ((int32_t)(U32(x))) +#define S16(x) ((int16_t)(U16(x))) +#define F32(x) (((__u32_f32_union__.u = U32(x)) & 0) + __u32_f32_union__.f) + +#define W32(x, v) \ + { \ + *((uint8_t*)x + 3) = ((v)&0xFF); \ + *((uint8_t*)x + 2) = (((v) >> 8) & 0xFF); \ + *((uint8_t*)x + 1) = (((v) >> 16) & 0xFF); \ + *((uint8_t*)x + 0) = (((v) >> 24) & 0xFF); \ + } + +#endif
\ No newline at end of file diff --git a/ZAPD/yaz0/yaz0.cpp b/ZAPD/yaz0/yaz0.cpp new file mode 100644 index 0000000..885f41d --- /dev/null +++ b/ZAPD/yaz0/yaz0.cpp @@ -0,0 +1,227 @@ +#include <stdint.h> +#include <stdio.h> +#include <stdlib.h> +#include <string.h> +#include <list> +#include "readwrite.h" + +#include "yaz0.h" + +typedef uint8_t u8; +typedef uint16_t u16; +typedef uint32_t u32; + +/* internal declarations */ +int yaz0_encode_internal(const u8* src, int srcSize, u8* Data); + +int yaz0_get_size(u8* src) { return U32(src + 0x4); } + +u32 toDWORD(u32 d) { + u8 w1 = d & 0xFF; + u8 w2 = (d >> 8) & 0xFF; + u8 w3 = (d >> 16) & 0xFF; + u8 w4 = d >> 24; + return (w1 << 24) | (w2 << 16) | (w3 << 8) | w4; +} + +// simple and straight encoding scheme for Yaz0 +u32 longest_match_brute(const u8* src, int size, int pos, u32* pMatchPos) { + int startPos = pos - 0x1000; + int max_match_size = size - pos; + u32 best_match_size = 0; + u32 best_match_pos = 0; + + if (max_match_size < 3) return 0; + + if (startPos < 0) startPos = 0; + + if (max_match_size > 0x111) max_match_size = 0x111; + + for (int i = startPos; i < pos; i++) { + int current_size; + for (current_size = 0; current_size < max_match_size; current_size++) { + if (src[i + current_size] != src[pos + current_size]) { + break; + } + } + if (current_size > best_match_size) { + best_match_size = current_size; + best_match_pos = i; + if (best_match_size == 0x111) break; + } + } + *pMatchPos = best_match_pos; + return best_match_size; +} + +u32 longest_match_rabinkarp(const u8* src, int size, int pos, u32* match_pos) { + int startPos = pos - 0x1000; + int max_match_size = size - pos; + u32 best_match_size = 0; + u32 best_match_pos = 0; + + if (max_match_size < 3) return 0; + + if (startPos < 0) startPos = 0; + + if (max_match_size > 0x111) max_match_size = 0x111; + + int find_hash = src[pos] << 16 | src[pos + 1] << 8 | src[pos + 2]; + int current_hash = src[startPos] << 16 | src[startPos + 1] << 8 | src[startPos + 2]; + + for (int i = startPos; i < pos; i++) { + if(current_hash == find_hash) { + int current_size; + for (current_size = 3; current_size < max_match_size; current_size++) { + if (src[i + current_size] != src[pos + current_size]) { + break; + } + } + if (current_size > best_match_size) { + best_match_size = current_size; + best_match_pos = i; + if (best_match_size == 0x111) break; + } + } + current_hash = (current_hash << 8 | src[i + 3]) & 0xFFFFFF; + } + *match_pos = best_match_pos; + + return best_match_size; +} + +int yaz0_encode_internal(const u8* src, int srcSize, u8* Data) { + int srcPos = 0; + + int bitmask = 0x80; + u8 currCodeByte = 0; + int currCodeBytePos = 0; + int pos = currCodeBytePos + 1; + + while (srcPos < srcSize) { + u32 numBytes; + u32 matchPos; + + numBytes = longest_match_rabinkarp(src, srcSize, srcPos, &matchPos); + //fprintf(stderr, "pos %x len %x pos %x\n", srcPos, (int)numBytes, (int)matchPos); + if (numBytes < 3) { + //fprintf(stderr, "single byte %02x\n", src[srcPos]); + Data[pos++] = src[srcPos++]; + currCodeByte |= bitmask; + } else { + // RLE part + u32 dist = srcPos - matchPos - 1; + + if (numBytes >= 0x12) // 3 byte encoding + { + Data[pos++] = dist >> 8; // 0R + Data[pos++] = dist & 0xFF; // FF + if (numBytes > 0xFF + 0x12) numBytes = 0xFF + 0x12; + Data[pos++] = numBytes - 0x12; + } else // 2 byte encoding + { + Data[pos++] = ((numBytes - 2) << 4) | (dist >> 8); + Data[pos++] = dist & 0xFF; + } + srcPos += numBytes; + } + bitmask >>= 1; + // write eight codes + if (!bitmask) { + Data[currCodeBytePos] = currCodeByte; + currCodeBytePos = pos++; + + currCodeByte = 0; + bitmask = 0x80; + } + } + if (bitmask) { + Data[currCodeBytePos] = currCodeByte; + } + + return pos; +} + +std::vector<uint8_t> yaz0_encode_fast(const u8* src, int src_size) { + std::vector<uint8_t> buffer; + std::vector<std::list<uint32_t>> lut; + lut.resize(0x1000000); + + for (int i = 0; i < src_size - 3; ++i) { + lut[src[i + 0] << 16 | src[i + 1] << 8 | src[i + 2]].push_back(i); + } + + return buffer; +} + +std::vector<uint8_t> yaz0_encode(const u8* src, int src_size) { + std::vector<uint8_t> buffer(src_size * 10 / 8 + 16); + u8* dst = buffer.data(); + + // write 4 bytes yaz0 header + memcpy(dst, "Yaz0", 4); + + // write 4 bytes uncompressed size + W32(dst + 4, src_size); + + // encode + int dst_size = yaz0_encode_internal(src, src_size, dst + 16); + int aligned_size = (dst_size + 31) & -16; + buffer.resize(aligned_size); + +#if 0 + std::vector<uint8_t> decompressed(src_size); + yaz0_decode(buffer.data(), decompressed.data(), src_size); + if(memcmp(src, decompressed.data(), src_size)) { + fprintf(stderr, "Decompressed buffer is different from original\n"); + } +#endif + + return buffer; +} + +void yaz0_decode(const uint8_t* source, uint8_t* decomp, int32_t decompSize) { + uint32_t srcPlace = 0, dstPlace = 0; + uint32_t i, dist, copyPlace, numBytes; + uint8_t codeByte, byte1, byte2; + uint8_t bitCount = 0; + + source += 0x10; + while (dstPlace < decompSize) { + /* If there are no more bits to test, get a new byte */ + if (!bitCount) { + codeByte = source[srcPlace++]; + bitCount = 8; + } + + /* If bit 7 is a 1, just copy 1 byte from source to destination */ + /* Else do some decoding */ + if (codeByte & 0x80) { + decomp[dstPlace++] = source[srcPlace++]; + } else { + /* Get 2 bytes from source */ + byte1 = source[srcPlace++]; + byte2 = source[srcPlace++]; + + /* Calculate distance to move in destination */ + /* And the number of bytes to copy */ + dist = ((byte1 & 0xF) << 8) | byte2; + copyPlace = dstPlace - (dist + 1); + numBytes = byte1 >> 4; + + /* Do more calculations on the number of bytes to copy */ + if (!numBytes) + numBytes = source[srcPlace++] + 0x12; + else + numBytes += 2; + + /* Copy data from a previous point in destination */ + /* to current point in destination */ + for (i = 0; i < numBytes; i++) decomp[dstPlace++] = decomp[copyPlace++]; + } + + /* Set up for the next read cycle */ + codeByte = codeByte << 1; + bitCount--; + } +}
\ No newline at end of file diff --git a/ZAPD/yaz0/yaz0.h b/ZAPD/yaz0/yaz0.h new file mode 100644 index 0000000..0d5cc3c --- /dev/null +++ b/ZAPD/yaz0/yaz0.h @@ -0,0 +1,6 @@ +#pragma once + +#include <vector> + +void yaz0_decode(const uint8_t* src, uint8_t* dest, int32_t destsize); +std::vector<uint8_t> yaz0_encode(const uint8_t* src, int src_size);
\ No newline at end of file |
