forked from mia/Aegisub
445 lines
9.8 KiB
C++
445 lines
9.8 KiB
C++
// Copyright (c) 2008, Niels Martin Hansen
|
|
//
|
|
// 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 of the Aegisub Group 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 THE COPYRIGHT HOLDERS AND CONTRIBUTORS "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 THE COPYRIGHT OWNER OR CONTRIBUTORS 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.
|
|
//
|
|
// Aegisub Project http://www.aegisub.org/
|
|
|
|
/// @file aegisub_endian.h
|
|
/// @brief Convert numbers between various endianness
|
|
/// @ingroup utility
|
|
///
|
|
|
|
|
|
// Sanity check
|
|
|
|
#ifndef HAVE_LITTLE_ENDIAN
|
|
# ifndef HAVE_BIG_ENDIAN
|
|
// We neither have big nor little endian from configuration
|
|
# ifdef HAVE_UNIVERSAL_ENDIAN
|
|
// But this is an OS X system building a universal binary
|
|
// Apple's GCC defines _BIG_ENDIAN when building for PPC
|
|
# ifdef _BIG_ENDIAN
|
|
|
|
/// DOCME
|
|
# define HAVE_BIG_ENDIAN
|
|
# else
|
|
|
|
/// DOCME
|
|
# define HAVE_LITTLE_ENDIAN
|
|
# endif
|
|
|
|
/// DOCME
|
|
# undef HAVE_DYNAMIC_ENDIAN
|
|
# else // !HAVE_UNIVERSAL_ENDIAN
|
|
// We aren't building an OS X universal binary
|
|
// Use the dynamic endian code
|
|
# ifndef HAVE_DYNAMIC_ENDIAN
|
|
|
|
/// DOCME
|
|
# define HAVE_DYNAMIC_ENDIAN
|
|
# endif
|
|
# endif //HAVE_UNIVERSAL_ENDIAN
|
|
# endif // HAVE_BIG_ENDIAN
|
|
#endif // HAVE_LITTLE_ENDIAN
|
|
|
|
#ifdef HAVE_LITTLE_ENDIAN
|
|
# ifdef HAVE_BIG_ENDIAN
|
|
# error You cannot have both HAVE_LITTLE_ENDIAN and HAVE_BIG_ENDIAN defined at the same time
|
|
# endif
|
|
#endif
|
|
|
|
#ifndef AGI_PRE
|
|
#include <cstdint>
|
|
#endif
|
|
|
|
|
|
/// DOCME
|
|
namespace Endian {
|
|
|
|
// Unconditionally reverse endianness
|
|
|
|
// These are only defined for unsigned ints,
|
|
// Use reinterpret_cast on the values if you need signed values.
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint16_t Reverse(uint16_t val)
|
|
{
|
|
return
|
|
((val & 0x00FF) << 8) |
|
|
((val & 0xFF00) >> 8);
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint32_t Reverse(uint32_t val)
|
|
{
|
|
return
|
|
((val & 0x000000FF) << 24) |
|
|
((val & 0x0000FF00) << 8) |
|
|
((val & 0x00FF0000) >> 8) |
|
|
((val & 0xFF000000) >> 24);
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint64_t Reverse(uint64_t val)
|
|
{
|
|
return
|
|
((val & 0x00000000000000FFULL) << 56) |
|
|
((val & 0x000000000000FF00ULL) << 40) |
|
|
((val & 0x0000000000FF0000ULL) << 24) |
|
|
((val & 0x00000000FF000000ULL) << 8) |
|
|
((val & 0x000000FF00000000ULL) >> 8) |
|
|
((val & 0x0000FF0000000000ULL) >> 24) |
|
|
((val & 0x00FF000000000000ULL) >> 40) |
|
|
((val & 0xFF00000000000000ULL) >> 56);
|
|
}
|
|
|
|
|
|
#ifndef HAVE_DYNAMIC_ENDIAN
|
|
// Regular, fast, templatized conditional reversing
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
template <class T>
|
|
inline T LittleToMachine(T val)
|
|
{
|
|
#ifdef HAVE_BIG_ENDIAN
|
|
// We're on big endian, reverse little to big
|
|
return Reverse(val);
|
|
#else
|
|
// We're on little endian and input is little
|
|
return val;
|
|
#endif
|
|
}
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
template <class T>
|
|
inline T BigToMachine(T val)
|
|
{
|
|
#ifdef HAVE_LITTLE_ENDIAN
|
|
// We're on little endian, reverse big to little
|
|
return Reverse(val);
|
|
#else
|
|
// We're on big endian and input is big
|
|
return val;
|
|
#endif
|
|
}
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
template <class T>
|
|
inline T MachineToLittle(T val)
|
|
{
|
|
#ifdef HAVE_BIG_ENDIAN
|
|
// We're on big endian, reverse to little
|
|
return Reverse(val);
|
|
#else
|
|
// Already on little, nothing to be done
|
|
return val;
|
|
#endif
|
|
}
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
template <class T>
|
|
inline T MachineToBig(T val)
|
|
{
|
|
#ifdef HAVE_LITTLE_ENDIAN
|
|
// We're on little endian, reverse to big
|
|
return Reverse(val);
|
|
#else
|
|
// Already on big, nothing to be done
|
|
return val;
|
|
#endif
|
|
}
|
|
|
|
#else // HAVE_DYNAMIC_ENDIAN
|
|
|
|
// Dynamic endianness handling
|
|
|
|
// Exploit that bit-shifting operations always can put bytes into
|
|
// machine word order, while unions can be used to access bytes
|
|
// only from an explicitly given byte order.
|
|
// This is probably slower than when we explicitly know
|
|
// the endianness of the machine we are on, but it's the same
|
|
// code for any platform!
|
|
|
|
// Unions to pack together ints and get their physical bytes
|
|
|
|
|
|
/// DOCME
|
|
union bytes16 {
|
|
/// DOCME
|
|
uint8_t byte[2];
|
|
|
|
/// DOCME
|
|
uint16_t word;
|
|
};
|
|
|
|
/// DOCME
|
|
union bytes32 {
|
|
/// DOCME
|
|
uint8_t byte[4];
|
|
|
|
/// DOCME
|
|
uint32_t word;
|
|
};
|
|
|
|
/// DOCME
|
|
union bytes64 {
|
|
/// DOCME
|
|
uint8_t byte[8];
|
|
|
|
/// DOCME
|
|
uint64_t word;
|
|
};
|
|
|
|
// 16 bit words
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint16_t MachineToBig(uint16_t val)
|
|
{
|
|
bytes16 pack;
|
|
// Store the bytes into the correct positions in the word
|
|
pack.byte[0] = (val & 0xFF00) >> 8;
|
|
pack.byte[1] = val & 0x00FF;
|
|
// And return a value now encoded as big endian
|
|
return pack.word;
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint16_t MachineToLittle(uint16_t val)
|
|
{
|
|
bytes16 pack;
|
|
// Store the bytes into the correct positions in the word
|
|
pack.byte[0] = val & 0x00FF;
|
|
pack.byte[1] = (val & 0xFF00) >> 8;
|
|
// And return a value now encoded as little endian
|
|
return pack.word;
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint16_t BigToMachine(uint16_t val)
|
|
{
|
|
bytes16 pack;
|
|
// Put our word into the pack
|
|
pack.word = val;
|
|
// And produce a machine endian value of it
|
|
return uint16_t(pack.byte[1]) | (uint16_t(pack.byte[0]) << 8);
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint16_t LittleToMachine(uint16_t val)
|
|
{
|
|
bytes16 pack;
|
|
// Put our word into the pack
|
|
pack.word = val;
|
|
// And produce a machine endian value of it
|
|
return uint16_t(pack.byte[0]) | (uint16_t(pack.byte[1]) << 8);
|
|
}
|
|
|
|
|
|
// 32 bit words
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint32_t MachineToBig(uint32_t val)
|
|
{
|
|
bytes32 pack;
|
|
pack.byte[0] = (val & 0xFF000000) >> 24;
|
|
pack.byte[1] = (val & 0x00FF0000) >> 16;
|
|
pack.byte[2] = (val & 0x0000FF00) >> 8;
|
|
pack.byte[3] = val & 0x000000FF ;
|
|
return pack.word;
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint32_t MachineToLittle(uint32_t val)
|
|
{
|
|
bytes32 pack;
|
|
pack.byte[0] = val & 0x000000FF ;
|
|
pack.byte[1] = (val & 0x0000FF00) >> 8;
|
|
pack.byte[2] = (val & 0x00FF0000) >> 16;
|
|
pack.byte[3] = (val & 0xFF000000) >> 24;
|
|
return pack.word;
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint32_t BigToMachine(uint32_t val)
|
|
{
|
|
bytes32 pack;
|
|
pack.word = val;
|
|
return
|
|
(uint32_t(pack.byte[0]) << 24) |
|
|
(uint32_t(pack.byte[1]) << 16) |
|
|
(uint32_t(pack.byte[2]) << 8) |
|
|
uint32_t(pack.byte[3]);
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint32_t LittleToMachine(uint32_t val)
|
|
{
|
|
bytes32 pack;
|
|
pack.word = val;
|
|
return
|
|
(uint32_t(pack.byte[3]) << 24) |
|
|
(uint32_t(pack.byte[2]) << 16) |
|
|
(uint32_t(pack.byte[1]) << 8) |
|
|
uint32_t(pack.byte[0]);
|
|
}
|
|
|
|
|
|
// 64 bit words
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint64_t MachineToBig(uint64_t val)
|
|
{
|
|
bytes64 pack;
|
|
pack.byte[0] = (val & 0xFF00000000000000ULL) >> 56;
|
|
pack.byte[1] = (val & 0x00FF000000000000ULL) >> 48;
|
|
pack.byte[2] = (val & 0x0000FF0000000000ULL) >> 40;
|
|
pack.byte[3] = (val & 0x000000FF00000000ULL) >> 32;
|
|
pack.byte[4] = (val & 0x00000000FF000000ULL) >> 24;
|
|
pack.byte[5] = (val & 0x0000000000FF0000ULL) >> 16;
|
|
pack.byte[6] = (val & 0x000000000000FF00ULL) >> 8;
|
|
pack.byte[7] = val & 0x00000000000000FFULL ;
|
|
return pack.word;
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint64_t MachineToLittle(uint64_t val)
|
|
{
|
|
bytes64 pack;
|
|
pack.byte[0] = val & 0x00000000000000FFULL ;
|
|
pack.byte[1] = (val & 0x000000000000FF00ULL) >> 8;
|
|
pack.byte[2] = (val & 0x0000000000FF0000ULL) >> 16;
|
|
pack.byte[3] = (val & 0x00000000FF000000ULL) >> 24;
|
|
pack.byte[4] = (val & 0x000000FF00000000ULL) >> 32;
|
|
pack.byte[5] = (val & 0x0000FF0000000000ULL) >> 40;
|
|
pack.byte[6] = (val & 0x00FF000000000000ULL) >> 48;
|
|
pack.byte[7] = (val & 0xFF00000000000000ULL) >> 56;
|
|
return pack.word;
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
/// @return
|
|
///
|
|
inline uint64_t BigToMachine(uint64_t val)
|
|
{
|
|
bytes64 pack;
|
|
pack.word = val;
|
|
return
|
|
(uint64_t(pack.byte[0]) << 56) |
|
|
(uint64_t(pack.byte[1]) << 48) |
|
|
(uint64_t(pack.byte[2]) << 40) |
|
|
(uint64_t(pack.byte[3]) << 32) |
|
|
(uint64_t(pack.byte[4]) << 24) |
|
|
(uint64_t(pack.byte[5]) << 16) |
|
|
(uint64_t(pack.byte[6]) << 8) |
|
|
uint64_t(pack.byte[7]);
|
|
}
|
|
|
|
|
|
/// @brief DOCME
|
|
/// @param val
|
|
///
|
|
inline uint64_t LittleToMachine(uint64_t val)
|
|
{
|
|
bytes64 pack;
|
|
pack.word = val;
|
|
return
|
|
(uint64_t(pack.byte[7]) << 56) |
|
|
(uint64_t(pack.byte[6]) << 48) |
|
|
(uint64_t(pack.byte[5]) << 40) |
|
|
(uint64_t(pack.byte[4]) << 32) |
|
|
(uint64_t(pack.byte[3]) << 24) |
|
|
(uint64_t(pack.byte[2]) << 16) |
|
|
(uint64_t(pack.byte[1]) << 8) |
|
|
uint64_t(pack.byte[0]);
|
|
}
|
|
|
|
#endif
|
|
}
|