This commit is contained in:
Chenwenxuan
2024-03-06 14:54:30 +08:00
commit edac2715f0
1525 changed files with 809982 additions and 0 deletions

View File

@@ -0,0 +1,166 @@
/**
* Copyright (C) 2015 Trustifier Inc.
* Copyright (C) 2015 Ahmed Masud
* Copyright (C) 2015 Topology LP
* Copyright (C) 2018 Jakob Petsovits
* All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*
* Adapted from https://github.com/ahmed-masud/libbase32,
* commit 79761b2b79b0545697945efe0987a8d3004512f9.
* Quite different now.
*/
#ifndef CPPCODEC_DETAIL_BASE32
#define CPPCODEC_DETAIL_BASE32
#include <stdint.h>
#include <stdlib.h> // for abort()
#include "../data/access.hpp"
#include "../parse_error.hpp"
#include "config.hpp"
#include "stream_codec.hpp"
namespace cppcodec {
namespace detail {
template <typename CodecVariant>
class base32 : public CodecVariant::template codec_impl<base32<CodecVariant>>
{
public:
static inline constexpr uint8_t binary_block_size() { return 5; }
static inline constexpr uint8_t encoded_block_size() { return 8; }
static CPPCODEC_ALWAYS_INLINE constexpr uint8_t num_encoded_tail_symbols(uint8_t num_bytes)
{
return (num_bytes == 1) ? 2 // 2 symbols, 6 padding characters
: (num_bytes == 2) ? 4 // 4 symbols, 4 padding characters
: (num_bytes == 3) ? 5 // 5 symbols, 3 padding characters
: (num_bytes == 4) ? 7 // 7 symbols, 1 padding characters
: throw std::domain_error("invalid number of bytes in a tail block");
}
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE constexpr uint8_t index(
const uint8_t* b /*binary block*/) noexcept
{
static_assert(I >= 0 && I < encoded_block_size(),
"invalid encoding symbol index in a block");
return (I == 0) ? ((b[0] >> 3) & 0x1F) // first 5 bits
: (I == 1) ? (((b[0] << 2) & 0x1C) | ((b[1] >> 6) & 0x3))
: (I == 2) ? ((b[1] >> 1) & 0x1F)
: (I == 3) ? (((b[1] << 4) & 0x10) | ((b[2] >> 4) & 0xF))
: (I == 4) ? (((b[2] << 1) & 0x1E) | ((b[3] >> 7) & 0x1))
: (I == 5) ? ((b[3] >> 2) & 0x1F)
: (I == 6) ? (((b[3] << 3) & 0x18) | ((b[4] >> 5) & 0x7))
: /*I == 7*/ (b[4] & 0x1F); // last 5 bits;
}
template <bool B>
using uint8_if = typename std::enable_if<B, uint8_t>::type;
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE constexpr
uint8_if<I == 1 || I == 3 || I == 4 || I == 6> index_last(
const uint8_t* b /*binary block*/) noexcept
{
return (I == 1) ? ((b[0] << 2) & 0x1C) // abbreviated 2nd symbol
: (I == 3) ? ((b[1] << 4) & 0x10) // abbreviated 4th symbol
: (I == 4) ? ((b[2] << 1) & 0x1E) // abbreviated 5th symbol
: /*I == 6*/ ((b[3] << 3) & 0x18); // abbreviated 7th symbol
}
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE
uint8_if<I != 1 && I != 3 && I != 4 && I != 6> index_last(
const uint8_t* /*binary block*/)
{
throw std::domain_error("invalid last encoding symbol index in a tail");
}
template <typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void decode_block(
Result& decoded, ResultState&, const alphabet_index_t* idx);
template <typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void decode_tail(
Result& decoded, ResultState&, const alphabet_index_t* idx, size_t idx_len);
};
//
// 11111111 10101010 10110011 10111100 10010100
// => 11111 11110 10101 01011 00111 01111 00100 10100
//
template <typename CodecVariant>
template <typename Result, typename ResultState>
CPPCODEC_ALWAYS_INLINE void base32<CodecVariant>::decode_block(
Result& decoded, ResultState& state, const alphabet_index_t* idx)
{
put(decoded, state, static_cast<uint8_t>(((idx[0] << 3) & 0xF8) | ((idx[1] >> 2) & 0x7)));
put(decoded, state, static_cast<uint8_t>(((idx[1] << 6) & 0xC0) | ((idx[2] << 1) & 0x3E) | ((idx[3] >> 4) & 0x1)));
put(decoded, state, static_cast<uint8_t>(((idx[3] << 4) & 0xF0) | ((idx[4] >> 1) & 0xF)));
put(decoded, state, static_cast<uint8_t>(((idx[4] << 7) & 0x80) | ((idx[5] << 2) & 0x7C) | ((idx[6] >> 3) & 0x3)));
put(decoded, state, static_cast<uint8_t>(((idx[6] << 5) & 0xE0) | (idx[7] & 0x1F)));
}
template <typename CodecVariant>
template <typename Result, typename ResultState>
CPPCODEC_ALWAYS_INLINE void base32<CodecVariant>::decode_tail(
Result& decoded, ResultState& state, const alphabet_index_t* idx, size_t idx_len)
{
if (idx_len == 1) {
throw invalid_input_length(
"invalid number of symbols in last base32 block: found 1, expected 2, 4, 5 or 7");
}
if (idx_len == 3) {
throw invalid_input_length(
"invalid number of symbols in last base32 block: found 3, expected 2, 4, 5 or 7");
}
if (idx_len == 6) {
throw invalid_input_length(
"invalid number of symbols in last base32 block: found 6, expected 2, 4, 5 or 7");
}
// idx_len == 2: decoded size 1
put(decoded, state, static_cast<uint8_t>(((idx[0] << 3) & 0xF8) | ((idx[1] >> 2) & 0x7)));
if (idx_len == 2) {
return;
}
// idx_len == 4: decoded size 2
put(decoded, state, static_cast<uint8_t>(((idx[1] << 6) & 0xC0) | ((idx[2] << 1) & 0x3E) | ((idx[3] >> 4) & 0x1)));
if (idx_len == 4) {
return;
}
// idx_len == 5: decoded size 3
put(decoded, state, static_cast<uint8_t>(((idx[3] << 4) & 0xF0) | ((idx[4] >> 1) & 0xF)));
if (idx_len == 5) {
return;
}
// idx_len == 7: decoded size 4
put(decoded, state, static_cast<uint8_t>(((idx[4] << 7) & 0x80) | ((idx[5] << 2) & 0x7C) | ((idx[6] >> 3) & 0x3)));
}
} // namespace detail
} // namespace cppcodec
#endif // CPPCODEC_DETAIL_BASE32

View File

@@ -0,0 +1,132 @@
/**
* Copyright (C) 2015 Topology LP
* Copyright (C) 2013 Adam Rudd (bit calculations)
* Copyright (C) 2018 Jakob Petsovits
* All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*
* Bit calculations adapted from https://github.com/adamvr/arduino-base64,
* commit 999595783185a0afcba156d7276dfeaa9cb5382f.
*/
#ifndef CPPCODEC_DETAIL_BASE64
#define CPPCODEC_DETAIL_BASE64
#include <stdexcept>
#include <stdint.h>
#include "../data/access.hpp"
#include "../parse_error.hpp"
#include "config.hpp"
#include "stream_codec.hpp"
namespace cppcodec {
namespace detail {
template <typename CodecVariant>
class base64 : public CodecVariant::template codec_impl<base64<CodecVariant>>
{
public:
static inline constexpr uint8_t binary_block_size() { return 3; }
static inline constexpr uint8_t encoded_block_size() { return 4; }
static CPPCODEC_ALWAYS_INLINE constexpr uint8_t num_encoded_tail_symbols(uint8_t num_bytes)
{
return (num_bytes == 1) ? 2 // 2 symbols, 2 padding characters
: (num_bytes == 2) ? 3 // 3 symbols, 1 padding character
: throw std::domain_error("invalid number of bytes in a tail block");
}
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE constexpr uint8_t index(
const uint8_t* b /*binary block*/) noexcept
{
static_assert(I >= 0 && I < encoded_block_size(),
"invalid encoding symbol index in a block");
return (I == 0) ? (b[0] >> 2) // first 6 bits
: (I == 1) ? (((b[0] & 0x3) << 4) | (b[1] >> 4))
: (I == 2) ? (((b[1] & 0xF) << 2) | (b[2] >> 6))
: /*I == 3*/ (b[2] & 0x3F); // last 6 bits
}
template <bool B>
using uint8_if = typename std::enable_if<B, uint8_t>::type;
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE constexpr uint8_if<I == 1 || I == 2> index_last(
const uint8_t* b /*binary block*/) noexcept
{
return (I == 1) ? ((b[0] & 0x3) << 4) // abbreviated 2nd symbol
: /*I == 2*/ ((b[1] & 0xF) << 2); // abbreviated 3rd symbol
}
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE uint8_if<I != 1 && I != 2> index_last(
const uint8_t* /*binary block*/)
{
throw std::domain_error("invalid last encoding symbol index in a tail");
}
template <typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void decode_block(
Result& decoded, ResultState&, const alphabet_index_t* idx);
template <typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void decode_tail(
Result& decoded, ResultState&, const alphabet_index_t* idx, size_t idx_len);
};
template <typename CodecVariant>
template <typename Result, typename ResultState>
CPPCODEC_ALWAYS_INLINE void base64<CodecVariant>::decode_block(
Result& decoded, ResultState& state, const alphabet_index_t* idx)
{
uint_fast32_t dec = (idx[0] << 18) | (idx[1] << 12) | (idx[2] << 6) | idx[3];
data::put(decoded, state, static_cast<uint8_t>(dec >> 16));
data::put(decoded, state, static_cast<uint8_t>((dec >> 8) & 0xFF));
data::put(decoded, state, static_cast<uint8_t>(dec & 0xFF));
}
template <typename CodecVariant>
template <typename Result, typename ResultState>
CPPCODEC_ALWAYS_INLINE void base64<CodecVariant>::decode_tail(
Result& decoded, ResultState& state, const alphabet_index_t* idx, size_t idx_len)
{
if (idx_len == 1) {
throw invalid_input_length(
"invalid number of symbols in last base64 block: found 1, expected 2 or 3");
}
// idx_len == 2: decoded size 1
data::put(decoded, state, static_cast<uint8_t>((idx[0] << 2) + ((idx[1] & 0x30) >> 4)));
if (idx_len == 2) {
return;
}
// idx_len == 3: decoded size 2
data::put(decoded, state, static_cast<uint8_t>(((idx[1] & 0xF) << 4) + ((idx[2] & 0x3C) >> 2)));
}
} // namespace detail
} // namespace cppcodec
#endif // CPPCODEC_DETAIL_BASE64

View File

@@ -0,0 +1,327 @@
/**
* Copyright (C) 2015 Topology LP
* All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#ifndef CPPCODEC_DETAIL_CODEC
#define CPPCODEC_DETAIL_CODEC
#include <assert.h>
#include <stdint.h>
#include <string>
#include <vector>
#include "../data/access.hpp"
#include "../data/raw_result_buffer.hpp"
namespace cppcodec {
namespace detail {
// SFINAE: Templates sometimes beat sensible overloads - make sure we don't call the wrong one.
template <typename T>
struct non_numeric : std::enable_if<!std::is_arithmetic<T>::value> { };
/**
* Public interface for all the codecs. For API documentation, see README.md.
*/
template <typename CodecImpl>
class codec
{
public:
//
// Encoding
// Convenient version, returns an std::string.
static std::string encode(const uint8_t* binary, size_t binary_size);
static std::string encode(const char* binary, size_t binary_size);
// static std::string encode(const T& binary); -> provided by template below
// Convenient version with templated result type.
template <typename Result> static Result encode(const uint8_t* binary, size_t binary_size);
template <typename Result> static Result encode(const char* binary, size_t binary_size);
template <typename Result = std::string, typename T = std::vector<uint8_t>>
static Result encode(const T& binary);
// Reused result container version. Resizes encoded_result before writing to it.
template <typename Result>
static void encode(Result& encoded_result, const uint8_t* binary, size_t binary_size);
template <typename Result>
static void encode(Result& encoded_result, const char* binary, size_t binary_size);
template <typename Result, typename T, typename non_numeric<T>::type* = nullptr>
static void encode(Result& encoded_result, const T& binary);
// Raw pointer output, assumes pre-allocated memory with size > encoded_size(binary_size).
static size_t encode(
char* encoded_result, size_t encoded_buffer_size,
const uint8_t* binary, size_t binary_size) noexcept;
static size_t encode(
char* encoded_result, size_t encoded_buffer_size,
const char* binary, size_t binary_size) noexcept;
template<typename T>
static size_t encode(
char* encoded_result, size_t encoded_buffer_size,
const T& binary) noexcept;
// Calculate the exact length of the encoded string based on binary size.
static constexpr size_t encoded_size(size_t binary_size) noexcept;
//
// Decoding
// Convenient version, returns an std::vector<uint8_t>.
static std::vector<uint8_t> decode(const char* encoded, size_t encoded_size);
// static std::vector<uint8_t> decode(const T& encoded); -> provided by template below
// Convenient version with templated result type.
template <typename Result> static Result decode(const char* encoded, size_t encoded_size);
template <typename Result = std::vector<uint8_t>, typename T = std::string>
static Result decode(const T& encoded);
// Reused result container version. Resizes binary_result before writing to it.
template <typename Result>
static void decode(Result& binary_result, const char* encoded, size_t encoded_size);
template <typename Result, typename T, typename non_numeric<T>::type* = nullptr>
static void decode(Result& binary_result, const T& encoded);
// Raw pointer output, assumes pre-allocated memory with size > decoded_max_size(encoded_size).
static size_t decode(
uint8_t* binary_result, size_t binary_buffer_size,
const char* encoded, size_t encoded_size);
static size_t decode(
char* binary_result, size_t binary_buffer_size,
const char* encoded, size_t encoded_size);
template<typename T> static size_t decode(
uint8_t* binary_result, size_t binary_buffer_size, const T& encoded);
template<typename T> static size_t decode(
char* binary_result, size_t binary_buffer_size, const T& encoded);
// Calculate the maximum size of the decoded binary buffer based on the encoded string length.
static constexpr size_t decoded_max_size(size_t encoded_size) noexcept;
};
//
// Inline definitions of the above functions, using CRTP to call into CodecImpl
//
//
// Encoding
template <typename CodecImpl>
inline std::string codec<CodecImpl>::encode(const uint8_t* binary, size_t binary_size)
{
return encode<std::string>(binary, binary_size);
}
template <typename CodecImpl>
inline std::string codec<CodecImpl>::encode(const char* binary, size_t binary_size)
{
return encode<std::string>(reinterpret_cast<const uint8_t*>(binary), binary_size);
}
template <typename CodecImpl>
template <typename Result>
inline Result codec<CodecImpl>::encode(const uint8_t* binary, size_t binary_size)
{
Result encoded_result;
encode(encoded_result, binary, binary_size);
return encoded_result;
}
template <typename CodecImpl>
template <typename Result>
inline Result codec<CodecImpl>::encode(const char* binary, size_t binary_size)
{
return encode<Result>(reinterpret_cast<const uint8_t*>(binary), binary_size);
}
template <typename CodecImpl>
template <typename Result, typename T>
inline Result codec<CodecImpl>::encode(const T& binary)
{
return encode<Result>(data::uchar_data(binary), data::size(binary));
}
template <typename CodecImpl>
template <typename Result>
inline void codec<CodecImpl>::encode(
Result& encoded_result, const uint8_t* binary, size_t binary_size)
{
// This overload is where we reserve buffer capacity and call into CodecImpl.
size_t encoded_buffer_size = encoded_size(binary_size);
auto state = data::create_state(encoded_result, data::specific_t());
data::init(encoded_result, state, encoded_buffer_size);
CodecImpl::encode(encoded_result, state, binary, binary_size);
data::finish(encoded_result, state);
assert(data::size(encoded_result) == encoded_buffer_size);
}
template <typename CodecImpl>
template <typename Result>
inline void codec<CodecImpl>::encode(
Result& encoded_result, const char* binary, size_t binary_size)
{
encode(encoded_result, reinterpret_cast<const uint8_t*>(binary), binary_size);
}
template <typename CodecImpl>
template <typename Result, typename T, typename non_numeric<T>::type*>
inline void codec<CodecImpl>::encode(Result& encoded_result, const T& binary)
{
encode(encoded_result, data::uchar_data(binary), data::size(binary));
}
template <typename CodecImpl>
inline size_t codec<CodecImpl>::encode(
char* encoded_result, size_t encoded_buffer_size,
const uint8_t* binary, size_t binary_size) noexcept
{
// This overload is where we wrap the result pointer & size.
data::raw_result_buffer encoded(encoded_result, encoded_buffer_size);
encode(encoded, binary, binary_size);
size_t encoded_size = data::size(encoded);
if (encoded_size < encoded_buffer_size) {
encoded_result[encoded_size] = '\0';
}
return encoded_size;
}
template <typename CodecImpl>
inline size_t codec<CodecImpl>::encode(
char* encoded_result, size_t encoded_buffer_size,
const char* binary, size_t binary_size) noexcept
{
// This overload is where we wrap the result pointer & size.
return encode(encoded_result, encoded_buffer_size,
reinterpret_cast<const uint8_t*>(binary), binary_size);
}
template <typename CodecImpl>
template <typename T>
inline size_t codec<CodecImpl>::encode(
char* encoded_result, size_t encoded_buffer_size,
const T& binary) noexcept
{
return encode(encoded_result, encoded_buffer_size, data::uchar_data(binary), data::size(binary));
}
template <typename CodecImpl>
inline constexpr size_t codec<CodecImpl>::encoded_size(size_t binary_size) noexcept
{
return CodecImpl::encoded_size(binary_size);
}
//
// Decoding
template <typename CodecImpl>
inline std::vector<uint8_t> codec<CodecImpl>::decode(const char* encoded, size_t encoded_size)
{
return decode<std::vector<uint8_t>>(encoded, encoded_size);
}
template <typename CodecImpl>
template <typename Result>
inline Result codec<CodecImpl>::decode(const char* encoded, size_t encoded_size)
{
Result result;
decode(result, encoded, encoded_size);
return result;
}
template <typename CodecImpl>
template <typename Result, typename T>
inline Result codec<CodecImpl>::decode(const T& encoded)
{
return decode<Result>(data::char_data(encoded), data::size(encoded));
}
template <typename CodecImpl>
template <typename Result>
inline void codec<CodecImpl>::decode(Result& binary_result, const char* encoded, size_t encoded_size)
{
// This overload is where we reserve buffer capacity and call into CodecImpl.
size_t binary_buffer_size = decoded_max_size(encoded_size);
auto state = data::create_state(binary_result, data::specific_t());
data::init(binary_result, state, binary_buffer_size);
CodecImpl::decode(binary_result, state, encoded, encoded_size);
data::finish(binary_result, state);
assert(data::size(binary_result) <= binary_buffer_size);
}
template <typename CodecImpl>
template <typename Result, typename T, typename non_numeric<T>::type*>
inline void codec<CodecImpl>::decode(Result& binary_result, const T& encoded)
{
decode(binary_result, data::char_data(encoded), data::size(encoded));
}
template <typename CodecImpl>
inline size_t codec<CodecImpl>::decode(
uint8_t* binary_result, size_t binary_buffer_size,
const char* encoded, size_t encoded_size)
{
return decode(reinterpret_cast<char*>(binary_result), binary_buffer_size, encoded, encoded_size);
}
template <typename CodecImpl>
inline size_t codec<CodecImpl>::decode(
char* binary_result, size_t binary_buffer_size,
const char* encoded, size_t encoded_size)
{
// This overload is where we wrap the result pointer & size.
data::raw_result_buffer binary(binary_result, binary_buffer_size);
decode(binary, encoded, encoded_size);
return data::size(binary);
}
template <typename CodecImpl>
template <typename T>
inline size_t codec<CodecImpl>::decode(
uint8_t* binary_result, size_t binary_buffer_size, const T& encoded)
{
return decode(reinterpret_cast<char*>(binary_result), binary_buffer_size, encoded);
}
template <typename CodecImpl>
template <typename T>
inline size_t codec<CodecImpl>::decode(char* binary_result, size_t binary_buffer_size, const T& encoded)
{
return decode(binary_result, binary_buffer_size, data::char_data(encoded), data::size(encoded));
}
template <typename CodecImpl>
inline constexpr size_t codec<CodecImpl>::decoded_max_size(size_t encoded_size) noexcept
{
return CodecImpl::decoded_max_size(encoded_size);
}
} // namespace detail
} // namespace cppcodec
#endif

View File

@@ -0,0 +1,40 @@
/**
* Copyright (C) 2015 Topology LP
* All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#ifndef CPPCODEC_DETAIL_CONFIG_HPP
#define CPPCODEC_DETAIL_CONFIG_HPP
#ifndef __has_attribute
#define __has_attribute(x) 0
#endif
#if __GNUC__ || __has_attribute(always_inline)
#define CPPCODEC_ALWAYS_INLINE inline __attribute__((always_inline))
#elif defined(_MSC_VER) && !defined(__INTEL_COMPILER)
#define CPPCODEC_ALWAYS_INLINE inline __forceinline
#else
#define CPPCODEC_ALWAYS_INLINE inline
#endif
#endif // CPPCODEC_DETAIL_CONFIG_HPP

View File

@@ -0,0 +1,114 @@
/**
* Copyright (C) 2015 Topology LP
* Copyright (C) 2018 Jakob Petsovits
* All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#ifndef CPPCODEC_DETAIL_HEX
#define CPPCODEC_DETAIL_HEX
#include <stdint.h>
#include <stdlib.h> // for abort()
#include "../data/access.hpp"
#include "../parse_error.hpp"
#include "stream_codec.hpp"
namespace cppcodec {
namespace detail {
template <typename CodecVariant>
class hex : public CodecVariant::template codec_impl<hex<CodecVariant>>
{
public:
static inline constexpr uint8_t binary_block_size() { return 1; }
static inline constexpr uint8_t encoded_block_size() { return 2; }
static CPPCODEC_ALWAYS_INLINE constexpr uint8_t num_encoded_tail_symbols(uint8_t /*num_bytes*/) noexcept
{
// Hex encoding only works on full bytes so there are no tails,
// no padding characters, and this function should (must) never be called.
return 0;
}
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE constexpr uint8_t index(
const uint8_t* b /*binary block*/) noexcept
{
static_assert(I >= 0 && I < encoded_block_size(),
"invalid encoding symbol index in a block");
return (I == 0) ? (b[0] >> 4) // first 4 bits
: /*I == 1*/ (b[0] & 0xF); // last 4 bits
}
// With only 2 bytes, enc<1> will always result in a full index() call and
// enc<0> will be protected by a not-reached assertion, so we don't actually
// care about index_last() except optimizing it out as good as possible.
template <bool B>
using uint8_if = typename std::enable_if<B, uint8_t>::type;
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE constexpr uint8_if<I == 0> index_last(
const uint8_t* /*binary block*/) noexcept
{
return 0;
}
template <uint8_t I>
static CPPCODEC_ALWAYS_INLINE uint8_if<I != 0> index_last(
const uint8_t* /*binary block*/)
{
throw std::domain_error("invalid last encoding symbol index in a tail");
}
template <typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void decode_block(
Result& decoded, ResultState&, const alphabet_index_t* idx);
template <typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void decode_tail(
Result& decoded, ResultState&, const alphabet_index_t* idx, size_t idx_len);
};
template <typename CodecVariant>
template <typename Result, typename ResultState>
CPPCODEC_ALWAYS_INLINE void hex<CodecVariant>::decode_block(
Result& decoded, ResultState& state, const alphabet_index_t* idx)
{
data::put(decoded, state, static_cast<uint8_t>((idx[0] << 4) | idx[1]));
}
template <typename CodecVariant>
template <typename Result, typename ResultState>
CPPCODEC_ALWAYS_INLINE void hex<CodecVariant>::decode_tail(
Result&, ResultState&, const alphabet_index_t*, size_t)
{
throw invalid_input_length(
"odd-length hex input is not supported by the streaming octet decoder, "
"use a place-based number decoder instead");
}
} // namespace detail
} // namespace cppcodec
#endif // CPPCODEC_DETAIL_HEX

View File

@@ -0,0 +1,439 @@
/**
* Copyright (C) 2015 Topology LP
* Copyright (C) 2018 Jakob Petsovits
* All rights reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#ifndef CPPCODEC_DETAIL_STREAM_CODEC
#define CPPCODEC_DETAIL_STREAM_CODEC
#include <limits>
#include <stdlib.h> // for abort()
#include <stdint.h>
#include "../parse_error.hpp"
#include "config.hpp"
namespace cppcodec {
namespace detail {
using alphabet_index_t = uint_fast16_t;
template <typename Codec, typename CodecVariant>
class stream_codec
{
public:
template <typename Result, typename ResultState> static void encode(
Result& encoded_result, ResultState&, const uint8_t* binary, size_t binary_size);
template <typename Result, typename ResultState> static void decode(
Result& binary_result, ResultState&, const char* encoded, size_t encoded_size);
static constexpr size_t encoded_size(size_t binary_size) noexcept;
static constexpr size_t decoded_max_size(size_t encoded_size) noexcept;
};
template <bool GeneratesPadding> // default for CodecVariant::generates_padding() == false
struct padder {
template <typename CodecVariant, typename Result, typename ResultState, typename SizeT>
static CPPCODEC_ALWAYS_INLINE void pad(Result&, ResultState&, SizeT) { }
};
template<> // specialization for CodecVariant::generates_padding() == true
struct padder<true> {
template <typename CodecVariant, typename Result, typename ResultState, typename SizeT>
static CPPCODEC_ALWAYS_INLINE void pad(
Result& encoded, ResultState& state, SizeT num_padding_characters)
{
for (SizeT i = 0; i < num_padding_characters; ++i) {
data::put(encoded, state, CodecVariant::padding_symbol());
}
}
};
template <size_t I>
struct enc {
// Block encoding: Go from 0 to (block size - 1), append a symbol for each iteration unconditionally.
template <typename Codec, typename CodecVariant, typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void block(Result& encoded, ResultState& state, const uint8_t* src)
{
using EncodedBlockSizeT = decltype(Codec::encoded_block_size());
constexpr static const EncodedBlockSizeT SymbolIndex = static_cast<EncodedBlockSizeT>(I - 1);
enc<I - 1>().template block<Codec, CodecVariant>(encoded, state, src);
data::put(encoded, state, CodecVariant::symbol(Codec::template index<SymbolIndex>(src)));
}
// Tail encoding: Go from 0 until (runtime) num_symbols, append a symbol for each iteration.
template <typename Codec, typename CodecVariant, typename Result, typename ResultState,
typename EncodedBlockSizeT = decltype(Codec::encoded_block_size())>
static CPPCODEC_ALWAYS_INLINE void tail(
Result& encoded, ResultState& state, const uint8_t* src, EncodedBlockSizeT num_symbols)
{
constexpr static const EncodedBlockSizeT SymbolIndex = Codec::encoded_block_size() - I;
constexpr static const EncodedBlockSizeT NumSymbols = SymbolIndex + static_cast<EncodedBlockSizeT>(1);
if (num_symbols == NumSymbols) {
data::put(encoded, state, CodecVariant::symbol(Codec::template index_last<SymbolIndex>(src)));
return;
}
data::put(encoded, state, CodecVariant::symbol(Codec::template index<SymbolIndex>(src)));
enc<I - 1>().template tail<Codec, CodecVariant>(encoded, state, src, num_symbols);
}
};
template<> // terminating specialization
struct enc<0> {
template <typename Codec, typename CodecVariant, typename Result, typename ResultState>
static CPPCODEC_ALWAYS_INLINE void block(Result&, ResultState&, const uint8_t*) { }
template <typename Codec, typename CodecVariant, typename Result, typename ResultState,
typename EncodedBlockSizeT = decltype(Codec::encoded_block_size())>
static CPPCODEC_ALWAYS_INLINE void tail(Result&, ResultState&, const uint8_t*, EncodedBlockSizeT)
{
abort(); // Not reached: block() should be called if num_symbols == block size, not tail().
}
};
template <typename Codec, typename CodecVariant>
template <typename Result, typename ResultState>
inline void stream_codec<Codec, CodecVariant>::encode(
Result& encoded_result, ResultState& state,
const uint8_t* src, size_t src_size)
{
using encoder = enc<Codec::encoded_block_size()>;
const uint8_t* src_end = src + src_size;
if (src_size >= Codec::binary_block_size()) {
src_end -= Codec::binary_block_size();
for (; src <= src_end; src += Codec::binary_block_size()) {
encoder::template block<Codec, CodecVariant>(encoded_result, state, src);
}
src_end += Codec::binary_block_size();
}
if (src_end > src) {
auto remaining_src_len = src_end - src;
if (!remaining_src_len || remaining_src_len >= Codec::binary_block_size()) {
abort();
return;
}
auto num_symbols = Codec::num_encoded_tail_symbols(
static_cast<uint8_t>(remaining_src_len));
encoder::template tail<Codec, CodecVariant>(encoded_result, state, src, num_symbols);
padder<CodecVariant::generates_padding()>::template pad<CodecVariant>(
encoded_result, state, Codec::encoded_block_size() - num_symbols);
}
}
// Range & lookup table generation, see
// http://stackoverflow.com/questions/13313980/populate-an-array-using-constexpr-at-compile-time
// and http://cplusadd.blogspot.ca/2013/02/c11-compile-time-lookup-tablearray-with.html
template<unsigned... Is> struct seq {};
template<unsigned N, unsigned... Is>
struct gen_seq : gen_seq<N - 4, N - 4, N - 3, N - 2, N - 1, Is...> {
// Clang up to 3.6 has a limit of 256 for template recursion,
// so pass a few more symbols at once to make it work.
static_assert(N % 4 == 0, "I must be divisible by 4 to eventually end at 0");
};
template<unsigned... Is>
struct gen_seq<0, Is...> : seq<Is...> {};
template <size_t N>
struct lookup_table_t {
alphabet_index_t lookup[N];
static constexpr size_t size = N;
};
template<typename LambdaType, unsigned... Is>
constexpr lookup_table_t<sizeof...(Is)> make_lookup_table(seq<Is...>, LambdaType value_for_index) {
return { { value_for_index(Is)... } };
}
template<unsigned N, typename LambdaType>
constexpr lookup_table_t<N> make_lookup_table(LambdaType evalFunc) {
return make_lookup_table(gen_seq<N>(), evalFunc);
}
// CodecVariant::symbol() provides a symbol for an index.
// Use recursive templates to get the inverse lookup table for fast decoding.
template <typename T>
static CPPCODEC_ALWAYS_INLINE constexpr size_t num_possible_values()
{
return static_cast<size_t>(
static_cast<intmax_t>((std::numeric_limits<T>::max)())
- static_cast<intmax_t>((std::numeric_limits<T>::min)()) + 1);
}
template <typename CodecVariant, alphabet_index_t InvalidIdx, size_t I>
struct index_if_in_alphabet {
static CPPCODEC_ALWAYS_INLINE constexpr alphabet_index_t for_symbol(char symbol)
{
return (CodecVariant::symbol(
static_cast<alphabet_index_t>(CodecVariant::alphabet_size() - I)) == symbol)
? static_cast<alphabet_index_t>(CodecVariant::alphabet_size() - I)
: index_if_in_alphabet<CodecVariant, InvalidIdx, I - 1>::for_symbol(symbol);
}
};
template <typename CodecVariant, alphabet_index_t InvalidIdx>
struct index_if_in_alphabet<CodecVariant, InvalidIdx, 0> { // terminating specialization
static CPPCODEC_ALWAYS_INLINE constexpr alphabet_index_t for_symbol(char)
{
return InvalidIdx;
}
};
template <typename CodecVariant, size_t I>
struct padding_searcher {
static CPPCODEC_ALWAYS_INLINE constexpr bool exists_padding_symbol()
{
// Clang up to 3.6 has a limit of 256 for template recursion,
// so pass a few more symbols at once to make it work.
static_assert(I % 4 == 0, "I must be divisible by 4 to eventually end at 0");
return CodecVariant::is_padding_symbol(
static_cast<char>(num_possible_values<char>() - I - 4))
|| CodecVariant::is_padding_symbol(
static_cast<char>(num_possible_values<char>() - I - 3))
|| CodecVariant::is_padding_symbol(
static_cast<char>(num_possible_values<char>() - I - 2))
|| CodecVariant::is_padding_symbol(
static_cast<char>(num_possible_values<char>() - I - 1))
|| padding_searcher<CodecVariant, I - 4>::exists_padding_symbol();
}
};
template <typename CodecVariant>
struct padding_searcher<CodecVariant, 0> { // terminating specialization
static CPPCODEC_ALWAYS_INLINE constexpr bool exists_padding_symbol() { return false; }
};
template <typename CodecVariant>
struct alphabet_index_info
{
static constexpr const size_t num_possible_symbols = num_possible_values<char>();
static constexpr const alphabet_index_t padding_idx = 1 << 8;
static constexpr const alphabet_index_t invalid_idx = 1 << 9;
static constexpr const alphabet_index_t eof_idx = 1 << 10;
static constexpr const alphabet_index_t stop_character_mask = static_cast<alphabet_index_t>(~0xFFu);
static constexpr const bool padding_allowed = padding_searcher<
CodecVariant, num_possible_symbols>::exists_padding_symbol();
static CPPCODEC_ALWAYS_INLINE constexpr bool allows_padding()
{
return padding_allowed;
}
static CPPCODEC_ALWAYS_INLINE constexpr bool is_padding(alphabet_index_t idx)
{
return allows_padding() ? (idx == padding_idx) : false;
}
static CPPCODEC_ALWAYS_INLINE constexpr bool is_invalid(alphabet_index_t idx) { return idx == invalid_idx; }
static CPPCODEC_ALWAYS_INLINE constexpr bool is_eof(alphabet_index_t idx) { return idx == eof_idx; }
static CPPCODEC_ALWAYS_INLINE constexpr bool is_stop_character(alphabet_index_t idx)
{
return (idx & stop_character_mask) != 0;
}
private:
static CPPCODEC_ALWAYS_INLINE constexpr
alphabet_index_t valid_index_or(alphabet_index_t a, alphabet_index_t b)
{
return a == invalid_idx ? b : a;
}
using idx_if_in_alphabet = index_if_in_alphabet<
CodecVariant, invalid_idx, CodecVariant::alphabet_size()>;
static CPPCODEC_ALWAYS_INLINE constexpr alphabet_index_t index_of(char symbol)
{
return valid_index_or(idx_if_in_alphabet::for_symbol(symbol),
CodecVariant::is_eof_symbol(symbol) ? eof_idx
: CodecVariant::is_padding_symbol(symbol) ? padding_idx
: invalid_idx);
}
// GCC <= 4.9 has a bug with retaining constexpr when passing a function pointer.
// To get around this, we'll create a callable with operator() and pass that one.
// Unfortunately, MSVC prior to VS 2017 (for MinSizeRel or Release builds)
// chokes on this by compiling the project in 20 minutes instead of seconds.
// So let's define two separate variants and remove the old GCC one whenever we
// decide not to support GCC < 5.0 anymore.
#if defined(__GNUC__) && !defined(__clang__) && __GNUC__ < 5
struct index_at {
CPPCODEC_ALWAYS_INLINE constexpr alphabet_index_t operator()(size_t symbol) const {
return index_of(CodecVariant::normalized_symbol(static_cast<char>(symbol)));
}
};
#else
static CPPCODEC_ALWAYS_INLINE constexpr alphabet_index_t index_at(size_t symbol)
{
return index_of(CodecVariant::normalized_symbol(static_cast<char>(symbol)));
}
#endif
public:
struct lookup {
static CPPCODEC_ALWAYS_INLINE alphabet_index_t for_symbol(char symbol)
{
#if defined(__GNUC__) && !defined(__clang__) && __GNUC__ < 5
static constexpr const auto t = make_lookup_table<num_possible_symbols>(index_at());
#else
static constexpr const auto t = make_lookup_table<num_possible_symbols>(&index_at);
#endif
static_assert(t.size == num_possible_symbols,
"lookup table must cover each possible (character) symbol");
return t.lookup[static_cast<uint8_t>(symbol)];
}
};
};
//
// At long last! The actual decode/encode functions.
template <typename Codec, typename CodecVariant>
template <typename Result, typename ResultState>
inline void stream_codec<Codec, CodecVariant>::decode(
Result& binary_result, ResultState& state,
const char* src_encoded, size_t src_size)
{
using alphabet_index_lookup = typename alphabet_index_info<CodecVariant>::lookup;
const char* src = src_encoded;
const char* src_end = src + src_size;
alphabet_index_t alphabet_indexes[Codec::encoded_block_size()] = {};
alphabet_indexes[0] = alphabet_index_info<CodecVariant>::eof_idx;
alphabet_index_t* const alphabet_index_start = &alphabet_indexes[0];
alphabet_index_t* const alphabet_index_end = &alphabet_indexes[Codec::encoded_block_size()];
alphabet_index_t* alphabet_index_ptr = &alphabet_indexes[0];
while (src < src_end) {
if (CodecVariant::should_ignore(*src)) {
++src;
continue;
}
*alphabet_index_ptr = alphabet_index_lookup::for_symbol(*src);
if (alphabet_index_info<CodecVariant>::is_stop_character(*alphabet_index_ptr)) {
break;
}
++src;
++alphabet_index_ptr;
if (alphabet_index_ptr == alphabet_index_end) {
Codec::decode_block(binary_result, state, alphabet_indexes);
alphabet_index_ptr = alphabet_index_start;
}
}
if (alphabet_index_info<CodecVariant>::is_invalid(*alphabet_index_ptr)) {
throw symbol_error(*src);
}
++src;
alphabet_index_t* last_index_ptr = alphabet_index_ptr;
if (alphabet_index_info<CodecVariant>::is_padding(*last_index_ptr)) {
if (last_index_ptr == alphabet_index_start) {
// Don't accept padding at the start of a block.
// The encoder should have omitted that padding altogether.
throw padding_error();
}
// We're in here because we just read a (first) padding character. Try to read more.
// Count with last_index_ptr, but store in alphabet_index_ptr so we don't
// overflow the array in case the input data is too long.
++last_index_ptr;
while (src < src_end) {
*alphabet_index_ptr = alphabet_index_lookup::for_symbol(*(src++));
if (alphabet_index_info<CodecVariant>::is_eof(*alphabet_index_ptr)) {
*alphabet_index_ptr = alphabet_index_info<CodecVariant>::padding_idx;
break;
}
if (!alphabet_index_info<CodecVariant>::is_padding(*alphabet_index_ptr)) {
throw padding_error();
}
++last_index_ptr;
if (last_index_ptr > alphabet_index_end) {
throw padding_error();
}
}
}
if (last_index_ptr != alphabet_index_start) {
if ((CodecVariant::requires_padding()
|| alphabet_index_info<CodecVariant>::is_padding(*alphabet_index_ptr)
) && last_index_ptr != alphabet_index_end)
{
// If the input is not a multiple of the block size then the input is incorrect.
throw padding_error();
}
if (alphabet_index_ptr >= alphabet_index_end) {
abort();
return;
}
Codec::decode_tail(binary_result, state, alphabet_indexes,
static_cast<size_t>(alphabet_index_ptr - alphabet_index_start));
}
}
template <typename Codec, typename CodecVariant>
inline constexpr size_t stream_codec<Codec, CodecVariant>::encoded_size(size_t binary_size) noexcept
{
using C = Codec;
// constexpr rules make this a lot harder to read than it actually is.
return CodecVariant::generates_padding()
// With padding, the encoded size is a multiple of the encoded block size.
// To calculate that, round the binary size up to multiple of the binary block size,
// then convert to encoded by multiplying with { base32: 8/5, base64: 4/3 }.
? (binary_size + (C::binary_block_size() - 1)
- ((binary_size + (C::binary_block_size() - 1)) % C::binary_block_size()))
* C::encoded_block_size() / C::binary_block_size()
// No padding: only pad to the next multiple of 5 bits, i.e. at most a single extra byte.
: (binary_size * C::encoded_block_size() / C::binary_block_size())
+ (((binary_size * C::encoded_block_size()) % C::binary_block_size()) ? 1 : 0);
}
template <typename Codec, typename CodecVariant>
inline constexpr size_t stream_codec<Codec, CodecVariant>::decoded_max_size(size_t encoded_size) noexcept
{
using C = Codec;
return CodecVariant::requires_padding()
? (encoded_size / C::encoded_block_size() * C::binary_block_size())
: (encoded_size / C::encoded_block_size() * C::binary_block_size())
+ ((encoded_size % C::encoded_block_size())
* C::binary_block_size() / C::encoded_block_size());
}
} // namespace detail
} // namespace cppcodec
#endif // CPPCODEC_DETAIL_STREAM_CODEC