Danbias/Code/Misc/Utilities/Packing/Packing.cpp

347 lines
6.7 KiB
C++
Raw Normal View History

#include "Packing.h"
/***************************
Packing
***************************/
#include <stdint.h>
namespace Oyster
{
namespace Packing
{
//bool (1-bit)
void Pack(unsigned char buffer[], bool i)
{
*buffer++ = i;
}
//char (8-bit)
void Pack(unsigned char buffer[], char i)
{
*buffer++ = i;
}
void Pack(unsigned char buffer[], unsigned char i)
{
*buffer++ = i;
}
//short (16-bit)
void Pack(unsigned char buffer[], short i)
{
*buffer++ = i >> 8;
*buffer++ = (char)i;
}
void Pack(unsigned char buffer[], unsigned short i)
{
*buffer++ = i >> 8;
*buffer++ = (char)i;
}
//int (32-bit)
void Pack(unsigned char buffer[], int i)
{
*buffer++ = i >> 24;
*buffer++ = i >> 16;
*buffer++ = i >> 8;
*buffer++ = i;
}
void Pack(unsigned char buffer[], unsigned int i)
{
*buffer++ = i >> 24;
*buffer++ = i >> 16;
*buffer++ = i >> 8;
*buffer++ = i;
}
//__int64 (64-bit)
void Pack(unsigned char buffer[], __int64 i)
{
*buffer++ = (char)(i >> 56);
*buffer++ = (char)(i >> 48);
*buffer++ = (char)(i >> 40);
*buffer++ = (char)(i >> 32);
*buffer++ = (char)(i >> 24);
*buffer++ = (char)(i >> 16);
*buffer++ = (char)(i >> 8);
*buffer++ = (char)i;
}
void Pack(unsigned char buffer[], unsigned __int64 i)
{
*buffer++ = (char)(i >> 56);
*buffer++ = (char)(i >> 48);
*buffer++ = (char)(i >> 40);
*buffer++ = (char)(i >> 32);
*buffer++ = (char)(i >> 24);
*buffer++ = (char)(i >> 16);
*buffer++ = (char)(i >> 8);
*buffer++ = (char)i;
}
//floating point (32, 64-bit)
void Pack(unsigned char buffer[], float i)
{
int tempFloat = (int)Pack754(i, 32, 8);
Pack(buffer, tempFloat);
}
void Pack(unsigned char buffer[], double i)
{
__int64 tempDouble = Pack754(i, 64, 11);
Pack(buffer, tempDouble);
}
//string
void Pack(unsigned char buffer[], char str[])
{
short len = (short)strlen(str);
Pack(buffer, len);
buffer += 2;
memcpy(buffer, str, len);
}
void Pack(unsigned char buffer[], std::string& str)
{
short len = (short)str.length();
Pack(buffer, len);
buffer += 2;
memcpy(buffer, str.c_str(), len);
}
unsigned __int64 Pack754(long double f, unsigned bits, unsigned expbits)
{
long double fnorm;
int shift;
long long sign, exp, significand;
unsigned significandbits = bits - expbits - 1; // -1 for sign bit
if (f == 0.0)
return 0; // get this special case out of the way
// check sign and begin normalization
if (f < 0)
{
sign = 1;
fnorm = -f;
}
else
{
sign = 0;
fnorm = f;
}
// get the normalized form of f and track the exponent
shift = 0;
while(fnorm >= 2.0)
{
fnorm /= 2.0;
shift++;
}
while(fnorm < 1.0)
{
fnorm *= 2.0;
shift--;
}
fnorm = fnorm - 1.0;
// calculate the binary form (non-float) of the significand data
significand = (long long)(fnorm * ((1LL << significandbits) + 0.5f));
// get the biased exponent
exp = shift + ((1 << (expbits - 1)) - 1); // shift + bias
// return the final answer
return (sign << (bits - 1)) | (exp << (bits - expbits - 1)) | significand;
}
/******************************
Unpacking
******************************/
//bool (1-bit)
bool Unpackb(unsigned char buffer[])
{
2014-02-15 09:29:54 +01:00
return (bool)(*buffer);
}
//char (8-bit)
char Unpackc(unsigned char buffer[])
{
if(*buffer <= 0x7f)
{
return *buffer;
}
else
{
return (-1 - (unsigned char)(0xffu - *buffer));
}
}
unsigned char UnpackC(unsigned char buffer[])
{
return *buffer;
}
//short (16-bit)
short Unpacks(unsigned char buffer[])
{
short i = ((short)buffer[0] << 8) | buffer[1];
if(i > 0x7fffu)
{
i = -1 - (unsigned short)(0xffffu - i);
}
return i;
}
unsigned short UnpackS(unsigned char buffer[])
{
return ((unsigned int)buffer[0] << 8) | buffer[1];
}
//int (32-bit)
int Unpacki(unsigned char buffer[])
{
int i = ((int)buffer[0] << 24) |
((int)buffer[1] << 16) |
((int)buffer[2] << 8) |
((int)buffer[3]);
if(i > 0x7fffffffu)
{
i = -1 - (int)(0xffffffffu - i);
}
return i;
}
unsigned int UnpackI(unsigned char buffer[])
{
return ((unsigned int)buffer[0] << 24) |
((unsigned int)buffer[1] << 16) |
((unsigned int)buffer[2] << 8) |
((unsigned int)buffer[3]);
}
//__int64 (64-bit)
__int64 Unpacki64(unsigned char buffer[])
{
__int64 i = ((__int64)buffer[0] << 56) |
((__int64)buffer[1] << 48) |
((__int64)buffer[2] << 40) |
((__int64)buffer[3] << 32) |
((__int64)buffer[4] << 24) |
((__int64)buffer[5] << 16) |
((__int64)buffer[6] << 8) |
(buffer[7]);
if(i > 0x7fffffffffffffffu)
{
i = -1 - (__int64)(0xffffffffffffffffu - i);
}
return i;
}
unsigned __int64 UnpackI64(unsigned char buffer[])
{
return ((__int64)buffer[0] << 56) |
((__int64)buffer[1] << 48) |
((__int64)buffer[2] << 40) |
((__int64)buffer[3] << 32) |
((__int64)buffer[4] << 24) |
((__int64)buffer[5] << 16) |
((__int64)buffer[6] << 8) |
((__int64)buffer[7]);
}
//floating point (32, 64-bit)
float Unpackf(unsigned char buffer[])
{
int tempFloat = Unpacki(buffer);
return (float)Unpack754(tempFloat, 32, 8);
}
double Unpackd(unsigned char buffer[])
{
__int64 tempDouble = Unpacki64(buffer);
return Unpack754(tempDouble, 64, 11);
}
//string
char* UnpackCStr(unsigned char buffer[])
{
short len = UnpackS(buffer);
char* str = new char[len+1];
buffer += 2;
for(int i = 0; i < len; i++)
{
str[i] = buffer[i];
}
str[len] = '\0';
return str;
}
std::string UnpackStr(unsigned char buffer[])
{
short len = UnpackS(buffer);
std::string temp;
temp.resize(len);
buffer += 2;
for(int i = 0; i < len; i++)
{
temp[i] = buffer[i];
}
return temp;
}
long double Unpack754(unsigned __int64 i, unsigned bits, unsigned expbits)
{
long double result;
long long shift;
unsigned bias;
unsigned significandbits = bits - expbits - 1; // -1 for sign bit
if (i == 0)
return 0.0;
// pull the significand
result = (long double)(i&((1LL << significandbits) - 1)); // mask
result /= (1LL << significandbits); // convert back to float
result += 1.0f; // add the one back on
// deal with the exponent
bias = (1 << (expbits - 1)) - 1;
shift = ((i >> significandbits) & ((1LL << expbits) - 1)) - bias;
while(shift > 0)
{
result *= 2.0;
shift--;
}
while(shift < 0)
{
result /= 2.0;
shift++;
}
// sign it
result *= (i >> (bits - 1)) & 1 ? -1.0 : 1.0;
return result;
}
}
}