From 276b9a93c2e3b9ac58a9ddb5fff4c8299c459222 Mon Sep 17 00:00:00 2001 From: Daniel Friesel Date: Mon, 17 Sep 2018 11:10:12 +0200 Subject: add nlohmann modenjson --- .../lib/modernjson/detail/output/binary_writer.hpp | 949 +++++++++++++++++++++ .../modernjson/detail/output/output_adapters.hpp | 113 +++ .../lib/modernjson/detail/output/serializer.hpp | 631 ++++++++++++++ 3 files changed, 1693 insertions(+) create mode 100644 include/lib/modernjson/detail/output/binary_writer.hpp create mode 100644 include/lib/modernjson/detail/output/output_adapters.hpp create mode 100644 include/lib/modernjson/detail/output/serializer.hpp (limited to 'include/lib/modernjson/detail/output') diff --git a/include/lib/modernjson/detail/output/binary_writer.hpp b/include/lib/modernjson/detail/output/binary_writer.hpp new file mode 100644 index 0000000..1726e0c --- /dev/null +++ b/include/lib/modernjson/detail/output/binary_writer.hpp @@ -0,0 +1,949 @@ +#pragma once + +#include // reverse +#include // array +#include // uint8_t, uint16_t, uint32_t, uint64_t +#include // memcpy +#include // numeric_limits + +#include +#include + +namespace nlohmann +{ +namespace detail +{ +/////////////////// +// binary writer // +/////////////////// + +/*! +@brief serialization to CBOR and MessagePack values +*/ +template +class binary_writer +{ + public: + /*! + @brief create a binary writer + + @param[in] adapter output adapter to write to + */ + explicit binary_writer(output_adapter_t adapter) : oa(adapter) + { + assert(oa); + } + + /*! + @brief[in] j JSON value to serialize + */ + void write_cbor(const BasicJsonType& j) + { + switch (j.type()) + { + case value_t::null: + { + oa->write_character(static_cast(0xF6)); + break; + } + + case value_t::boolean: + { + oa->write_character(j.m_value.boolean + ? static_cast(0xF5) + : static_cast(0xF4)); + break; + } + + case value_t::number_integer: + { + if (j.m_value.number_integer >= 0) + { + // CBOR does not differentiate between positive signed + // integers and unsigned integers. Therefore, we used the + // code from the value_t::number_unsigned case here. + if (j.m_value.number_integer <= 0x17) + { + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x18)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x19)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x1A)); + write_number(static_cast(j.m_value.number_integer)); + } + else + { + oa->write_character(static_cast(0x1B)); + write_number(static_cast(j.m_value.number_integer)); + } + } + else + { + // The conversions below encode the sign in the first + // byte, and the value is converted to a positive number. + const auto positive_number = -1 - j.m_value.number_integer; + if (j.m_value.number_integer >= -24) + { + write_number(static_cast(0x20 + positive_number)); + } + else if (positive_number <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x38)); + write_number(static_cast(positive_number)); + } + else if (positive_number <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x39)); + write_number(static_cast(positive_number)); + } + else if (positive_number <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x3A)); + write_number(static_cast(positive_number)); + } + else + { + oa->write_character(static_cast(0x3B)); + write_number(static_cast(positive_number)); + } + } + break; + } + + case value_t::number_unsigned: + { + if (j.m_value.number_unsigned <= 0x17) + { + write_number(static_cast(j.m_value.number_unsigned)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x18)); + write_number(static_cast(j.m_value.number_unsigned)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x19)); + write_number(static_cast(j.m_value.number_unsigned)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x1A)); + write_number(static_cast(j.m_value.number_unsigned)); + } + else + { + oa->write_character(static_cast(0x1B)); + write_number(static_cast(j.m_value.number_unsigned)); + } + break; + } + + case value_t::number_float: + { + oa->write_character(get_cbor_float_prefix(j.m_value.number_float)); + write_number(j.m_value.number_float); + break; + } + + case value_t::string: + { + // step 1: write control byte and the string length + const auto N = j.m_value.string->size(); + if (N <= 0x17) + { + write_number(static_cast(0x60 + N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x78)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x79)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x7A)); + write_number(static_cast(N)); + } + // LCOV_EXCL_START + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x7B)); + write_number(static_cast(N)); + } + // LCOV_EXCL_STOP + + // step 2: write the string + oa->write_characters( + reinterpret_cast(j.m_value.string->c_str()), + j.m_value.string->size()); + break; + } + + case value_t::array: + { + // step 1: write control byte and the array size + const auto N = j.m_value.array->size(); + if (N <= 0x17) + { + write_number(static_cast(0x80 + N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x98)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x99)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x9A)); + write_number(static_cast(N)); + } + // LCOV_EXCL_START + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0x9B)); + write_number(static_cast(N)); + } + // LCOV_EXCL_STOP + + // step 2: write each element + for (const auto& el : *j.m_value.array) + { + write_cbor(el); + } + break; + } + + case value_t::object: + { + // step 1: write control byte and the object size + const auto N = j.m_value.object->size(); + if (N <= 0x17) + { + write_number(static_cast(0xA0 + N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0xB8)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0xB9)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0xBA)); + write_number(static_cast(N)); + } + // LCOV_EXCL_START + else if (N <= (std::numeric_limits::max)()) + { + oa->write_character(static_cast(0xBB)); + write_number(static_cast(N)); + } + // LCOV_EXCL_STOP + + // step 2: write each element + for (const auto& el : *j.m_value.object) + { + write_cbor(el.first); + write_cbor(el.second); + } + break; + } + + default: + break; + } + } + + /*! + @brief[in] j JSON value to serialize + */ + void write_msgpack(const BasicJsonType& j) + { + switch (j.type()) + { + case value_t::null: // nil + { + oa->write_character(static_cast(0xC0)); + break; + } + + case value_t::boolean: // true and false + { + oa->write_character(j.m_value.boolean + ? static_cast(0xC3) + : static_cast(0xC2)); + break; + } + + case value_t::number_integer: + { + if (j.m_value.number_integer >= 0) + { + // MessagePack does not differentiate between positive + // signed integers and unsigned integers. Therefore, we used + // the code from the value_t::number_unsigned case here. + if (j.m_value.number_unsigned < 128) + { + // positive fixnum + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 8 + oa->write_character(static_cast(0xCC)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 16 + oa->write_character(static_cast(0xCD)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 32 + oa->write_character(static_cast(0xCE)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 64 + oa->write_character(static_cast(0xCF)); + write_number(static_cast(j.m_value.number_integer)); + } + } + else + { + if (j.m_value.number_integer >= -32) + { + // negative fixnum + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer >= (std::numeric_limits::min)() and + j.m_value.number_integer <= (std::numeric_limits::max)()) + { + // int 8 + oa->write_character(static_cast(0xD0)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer >= (std::numeric_limits::min)() and + j.m_value.number_integer <= (std::numeric_limits::max)()) + { + // int 16 + oa->write_character(static_cast(0xD1)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer >= (std::numeric_limits::min)() and + j.m_value.number_integer <= (std::numeric_limits::max)()) + { + // int 32 + oa->write_character(static_cast(0xD2)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_integer >= (std::numeric_limits::min)() and + j.m_value.number_integer <= (std::numeric_limits::max)()) + { + // int 64 + oa->write_character(static_cast(0xD3)); + write_number(static_cast(j.m_value.number_integer)); + } + } + break; + } + + case value_t::number_unsigned: + { + if (j.m_value.number_unsigned < 128) + { + // positive fixnum + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 8 + oa->write_character(static_cast(0xCC)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 16 + oa->write_character(static_cast(0xCD)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 32 + oa->write_character(static_cast(0xCE)); + write_number(static_cast(j.m_value.number_integer)); + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + // uint 64 + oa->write_character(static_cast(0xCF)); + write_number(static_cast(j.m_value.number_integer)); + } + break; + } + + case value_t::number_float: + { + oa->write_character(get_msgpack_float_prefix(j.m_value.number_float)); + write_number(j.m_value.number_float); + break; + } + + case value_t::string: + { + // step 1: write control byte and the string length + const auto N = j.m_value.string->size(); + if (N <= 31) + { + // fixstr + write_number(static_cast(0xA0 | N)); + } + else if (N <= (std::numeric_limits::max)()) + { + // str 8 + oa->write_character(static_cast(0xD9)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + // str 16 + oa->write_character(static_cast(0xDA)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + // str 32 + oa->write_character(static_cast(0xDB)); + write_number(static_cast(N)); + } + + // step 2: write the string + oa->write_characters( + reinterpret_cast(j.m_value.string->c_str()), + j.m_value.string->size()); + break; + } + + case value_t::array: + { + // step 1: write control byte and the array size + const auto N = j.m_value.array->size(); + if (N <= 15) + { + // fixarray + write_number(static_cast(0x90 | N)); + } + else if (N <= (std::numeric_limits::max)()) + { + // array 16 + oa->write_character(static_cast(0xDC)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + // array 32 + oa->write_character(static_cast(0xDD)); + write_number(static_cast(N)); + } + + // step 2: write each element + for (const auto& el : *j.m_value.array) + { + write_msgpack(el); + } + break; + } + + case value_t::object: + { + // step 1: write control byte and the object size + const auto N = j.m_value.object->size(); + if (N <= 15) + { + // fixmap + write_number(static_cast(0x80 | (N & 0xF))); + } + else if (N <= (std::numeric_limits::max)()) + { + // map 16 + oa->write_character(static_cast(0xDE)); + write_number(static_cast(N)); + } + else if (N <= (std::numeric_limits::max)()) + { + // map 32 + oa->write_character(static_cast(0xDF)); + write_number(static_cast(N)); + } + + // step 2: write each element + for (const auto& el : *j.m_value.object) + { + write_msgpack(el.first); + write_msgpack(el.second); + } + break; + } + + default: + break; + } + } + + /*! + @param[in] j JSON value to serialize + @param[in] use_count whether to use '#' prefixes (optimized format) + @param[in] use_type whether to use '$' prefixes (optimized format) + @param[in] add_prefix whether prefixes need to be used for this value + */ + void write_ubjson(const BasicJsonType& j, const bool use_count, + const bool use_type, const bool add_prefix = true) + { + switch (j.type()) + { + case value_t::null: + { + if (add_prefix) + { + oa->write_character(static_cast('Z')); + } + break; + } + + case value_t::boolean: + { + if (add_prefix) + oa->write_character(j.m_value.boolean + ? static_cast('T') + : static_cast('F')); + break; + } + + case value_t::number_integer: + { + write_number_with_ubjson_prefix(j.m_value.number_integer, add_prefix); + break; + } + + case value_t::number_unsigned: + { + write_number_with_ubjson_prefix(j.m_value.number_unsigned, add_prefix); + break; + } + + case value_t::number_float: + { + write_number_with_ubjson_prefix(j.m_value.number_float, add_prefix); + break; + } + + case value_t::string: + { + if (add_prefix) + { + oa->write_character(static_cast('S')); + } + write_number_with_ubjson_prefix(j.m_value.string->size(), true); + oa->write_characters( + reinterpret_cast(j.m_value.string->c_str()), + j.m_value.string->size()); + break; + } + + case value_t::array: + { + if (add_prefix) + { + oa->write_character(static_cast('[')); + } + + bool prefix_required = true; + if (use_type and not j.m_value.array->empty()) + { + assert(use_count); + const CharType first_prefix = ubjson_prefix(j.front()); + const bool same_prefix = std::all_of(j.begin() + 1, j.end(), + [this, first_prefix](const BasicJsonType & v) + { + return ubjson_prefix(v) == first_prefix; + }); + + if (same_prefix) + { + prefix_required = false; + oa->write_character(static_cast('$')); + oa->write_character(first_prefix); + } + } + + if (use_count) + { + oa->write_character(static_cast('#')); + write_number_with_ubjson_prefix(j.m_value.array->size(), true); + } + + for (const auto& el : *j.m_value.array) + { + write_ubjson(el, use_count, use_type, prefix_required); + } + + if (not use_count) + { + oa->write_character(static_cast(']')); + } + + break; + } + + case value_t::object: + { + if (add_prefix) + { + oa->write_character(static_cast('{')); + } + + bool prefix_required = true; + if (use_type and not j.m_value.object->empty()) + { + assert(use_count); + const CharType first_prefix = ubjson_prefix(j.front()); + const bool same_prefix = std::all_of(j.begin(), j.end(), + [this, first_prefix](const BasicJsonType & v) + { + return ubjson_prefix(v) == first_prefix; + }); + + if (same_prefix) + { + prefix_required = false; + oa->write_character(static_cast('$')); + oa->write_character(first_prefix); + } + } + + if (use_count) + { + oa->write_character(static_cast('#')); + write_number_with_ubjson_prefix(j.m_value.object->size(), true); + } + + for (const auto& el : *j.m_value.object) + { + write_number_with_ubjson_prefix(el.first.size(), true); + oa->write_characters( + reinterpret_cast(el.first.c_str()), + el.first.size()); + write_ubjson(el.second, use_count, use_type, prefix_required); + } + + if (not use_count) + { + oa->write_character(static_cast('}')); + } + + break; + } + + default: + break; + } + } + + private: + /* + @brief write a number to output input + + @param[in] n number of type @a NumberType + @tparam NumberType the type of the number + + @note This function needs to respect the system's endianess, because bytes + in CBOR, MessagePack, and UBJSON are stored in network order (big + endian) and therefore need reordering on little endian systems. + */ + template + void write_number(const NumberType n) + { + // step 1: write number to array of length NumberType + std::array vec; + std::memcpy(vec.data(), &n, sizeof(NumberType)); + + // step 2: write array to output (with possible reordering) + if (is_little_endian) + { + // reverse byte order prior to conversion if necessary + std::reverse(vec.begin(), vec.end()); + } + + oa->write_characters(vec.data(), sizeof(NumberType)); + } + + // UBJSON: write number (floating point) + template::value, int>::type = 0> + void write_number_with_ubjson_prefix(const NumberType n, + const bool add_prefix) + { + if (add_prefix) + { + oa->write_character(get_ubjson_float_prefix(n)); + } + write_number(n); + } + + // UBJSON: write number (unsigned integer) + template::value, int>::type = 0> + void write_number_with_ubjson_prefix(const NumberType n, + const bool add_prefix) + { + if (n <= static_cast((std::numeric_limits::max)())) + { + if (add_prefix) + { + oa->write_character(static_cast('i')); // int8 + } + write_number(static_cast(n)); + } + else if (n <= (std::numeric_limits::max)()) + { + if (add_prefix) + { + oa->write_character(static_cast('U')); // uint8 + } + write_number(static_cast(n)); + } + else if (n <= static_cast((std::numeric_limits::max)())) + { + if (add_prefix) + { + oa->write_character(static_cast('I')); // int16 + } + write_number(static_cast(n)); + } + else if (n <= static_cast((std::numeric_limits::max)())) + { + if (add_prefix) + { + oa->write_character(static_cast('l')); // int32 + } + write_number(static_cast(n)); + } + else if (n <= static_cast((std::numeric_limits::max)())) + { + if (add_prefix) + { + oa->write_character(static_cast('L')); // int64 + } + write_number(static_cast(n)); + } + else + { + JSON_THROW(out_of_range::create(407, "number overflow serializing " + std::to_string(n))); + } + } + + // UBJSON: write number (signed integer) + template::value and + not std::is_floating_point::value, int>::type = 0> + void write_number_with_ubjson_prefix(const NumberType n, + const bool add_prefix) + { + if ((std::numeric_limits::min)() <= n and n <= (std::numeric_limits::max)()) + { + if (add_prefix) + { + oa->write_character(static_cast('i')); // int8 + } + write_number(static_cast(n)); + } + else if (static_cast((std::numeric_limits::min)()) <= n and n <= static_cast((std::numeric_limits::max)())) + { + if (add_prefix) + { + oa->write_character(static_cast('U')); // uint8 + } + write_number(static_cast(n)); + } + else if ((std::numeric_limits::min)() <= n and n <= (std::numeric_limits::max)()) + { + if (add_prefix) + { + oa->write_character(static_cast('I')); // int16 + } + write_number(static_cast(n)); + } + else if ((std::numeric_limits::min)() <= n and n <= (std::numeric_limits::max)()) + { + if (add_prefix) + { + oa->write_character(static_cast('l')); // int32 + } + write_number(static_cast(n)); + } + else if ((std::numeric_limits::min)() <= n and n <= (std::numeric_limits::max)()) + { + if (add_prefix) + { + oa->write_character(static_cast('L')); // int64 + } + write_number(static_cast(n)); + } + // LCOV_EXCL_START + else + { + JSON_THROW(out_of_range::create(407, "number overflow serializing " + std::to_string(n))); + } + // LCOV_EXCL_STOP + } + + /*! + @brief determine the type prefix of container values + + @note This function does not need to be 100% accurate when it comes to + integer limits. In case a number exceeds the limits of int64_t, + this will be detected by a later call to function + write_number_with_ubjson_prefix. Therefore, we return 'L' for any + value that does not fit the previous limits. + */ + CharType ubjson_prefix(const BasicJsonType& j) const noexcept + { + switch (j.type()) + { + case value_t::null: + return 'Z'; + + case value_t::boolean: + return j.m_value.boolean ? 'T' : 'F'; + + case value_t::number_integer: + { + if ((std::numeric_limits::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits::max)()) + { + return 'i'; + } + else if ((std::numeric_limits::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits::max)()) + { + return 'U'; + } + else if ((std::numeric_limits::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits::max)()) + { + return 'I'; + } + else if ((std::numeric_limits::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits::max)()) + { + return 'l'; + } + else // no check and assume int64_t (see note above) + { + return 'L'; + } + } + + case value_t::number_unsigned: + { + if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + return 'i'; + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + return 'U'; + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + return 'I'; + } + else if (j.m_value.number_unsigned <= (std::numeric_limits::max)()) + { + return 'l'; + } + else // no check and assume int64_t (see note above) + { + return 'L'; + } + } + + case value_t::number_float: + return get_ubjson_float_prefix(j.m_value.number_float); + + case value_t::string: + return 'S'; + + case value_t::array: + return '['; + + case value_t::object: + return '{'; + + default: // discarded values + return 'N'; + } + } + + static constexpr CharType get_cbor_float_prefix(float) + { + return static_cast(0xFA); // Single-Precision Float + } + + static constexpr CharType get_cbor_float_prefix(double) + { + return static_cast(0xFB); // Double-Precision Float + } + + static constexpr CharType get_msgpack_float_prefix(float) + { + return static_cast(0xCA); // float 32 + } + + static constexpr CharType get_msgpack_float_prefix(double) + { + return static_cast(0xCB); // float 64 + } + + static constexpr CharType get_ubjson_float_prefix(float) + { + return 'd'; // float 32 + } + + static constexpr CharType get_ubjson_float_prefix(double) + { + return 'D'; // float 64 + } + + private: + /// whether we can assume little endianess + const bool is_little_endian = binary_reader::little_endianess(); + + /// the output + output_adapter_t oa = nullptr; +}; +} +} diff --git a/include/lib/modernjson/detail/output/output_adapters.hpp b/include/lib/modernjson/detail/output/output_adapters.hpp new file mode 100644 index 0000000..ff86a6e --- /dev/null +++ b/include/lib/modernjson/detail/output/output_adapters.hpp @@ -0,0 +1,113 @@ +#pragma once + +#include // copy +#include // size_t +#include // streamsize +#include // back_inserter +#include // shared_ptr, make_shared +#include // basic_ostream +#include // basic_string +#include // vector + +namespace nlohmann +{ +namespace detail +{ +/// abstract output adapter interface +template struct output_adapter_protocol +{ + virtual void write_character(CharType c) = 0; + virtual void write_characters(const CharType* s, std::size_t length) = 0; + virtual ~output_adapter_protocol() = default; +}; + +/// a type to simplify interfaces +template +using output_adapter_t = std::shared_ptr>; + +/// output adapter for byte vectors +template +class output_vector_adapter : public output_adapter_protocol +{ + public: + explicit output_vector_adapter(std::vector& vec) : v(vec) {} + + void write_character(CharType c) override + { + v.push_back(c); + } + + void write_characters(const CharType* s, std::size_t length) override + { + std::copy(s, s + length, std::back_inserter(v)); + } + + private: + std::vector& v; +}; + +/// output adapter for output streams +template +class output_stream_adapter : public output_adapter_protocol +{ + public: + explicit output_stream_adapter(std::basic_ostream& s) : stream(s) {} + + void write_character(CharType c) override + { + stream.put(c); + } + + void write_characters(const CharType* s, std::size_t length) override + { + stream.write(s, static_cast(length)); + } + + private: + std::basic_ostream& stream; +}; + +/// output adapter for basic_string +template> +class output_string_adapter : public output_adapter_protocol +{ + public: + explicit output_string_adapter(StringType& s) : str(s) {} + + void write_character(CharType c) override + { + str.push_back(c); + } + + void write_characters(const CharType* s, std::size_t length) override + { + str.append(s, length); + } + + private: + StringType& str; +}; + +template> +class output_adapter +{ + public: + output_adapter(std::vector& vec) + : oa(std::make_shared>(vec)) {} + + output_adapter(std::basic_ostream& s) + : oa(std::make_shared>(s)) {} + + output_adapter(StringType& s) + : oa(std::make_shared>(s)) {} + + operator output_adapter_t() + { + return oa; + } + + private: + output_adapter_t oa = nullptr; +}; +} +} diff --git a/include/lib/modernjson/detail/output/serializer.hpp b/include/lib/modernjson/detail/output/serializer.hpp new file mode 100644 index 0000000..7c6abb8 --- /dev/null +++ b/include/lib/modernjson/detail/output/serializer.hpp @@ -0,0 +1,631 @@ +#pragma once + +#include // reverse, remove, fill, find, none_of +#include // array +#include // assert +#include // and, or +#include // localeconv, lconv +#include // labs, isfinite, isnan, signbit +#include // size_t, ptrdiff_t +#include // uint8_t +#include // snprintf +#include // numeric_limits +#include // string +#include // is_same + +#include +#include +#include +#include +#include +#include + +namespace nlohmann +{ +namespace detail +{ +/////////////////// +// serialization // +/////////////////// + +template +class serializer +{ + using string_t = typename BasicJsonType::string_t; + using number_float_t = typename BasicJsonType::number_float_t; + using number_integer_t = typename BasicJsonType::number_integer_t; + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + static constexpr uint8_t UTF8_ACCEPT = 0; + static constexpr uint8_t UTF8_REJECT = 1; + + public: + /*! + @param[in] s output stream to serialize to + @param[in] ichar indentation character to use + */ + serializer(output_adapter_t s, const char ichar) + : o(std::move(s)), loc(std::localeconv()), + thousands_sep(loc->thousands_sep == nullptr ? '\0' : * (loc->thousands_sep)), + decimal_point(loc->decimal_point == nullptr ? '\0' : * (loc->decimal_point)), + indent_char(ichar), indent_string(512, indent_char) + {} + + // delete because of pointer members + serializer(const serializer&) = delete; + serializer& operator=(const serializer&) = delete; + + /*! + @brief internal implementation of the serialization function + + This function is called by the public member function dump and organizes + the serialization internally. The indentation level is propagated as + additional parameter. In case of arrays and objects, the function is + called recursively. + + - strings and object keys are escaped using `escape_string()` + - integer numbers are converted implicitly via `operator<<` + - floating-point numbers are converted to a string using `"%g"` format + + @param[in] val value to serialize + @param[in] pretty_print whether the output shall be pretty-printed + @param[in] indent_step the indent level + @param[in] current_indent the current indent level (only used internally) + */ + void dump(const BasicJsonType& val, const bool pretty_print, + const bool ensure_ascii, + const unsigned int indent_step, + const unsigned int current_indent = 0) + { + switch (val.m_type) + { + case value_t::object: + { + if (val.m_value.object->empty()) + { + o->write_characters("{}", 2); + return; + } + + if (pretty_print) + { + o->write_characters("{\n", 2); + + // variable to hold indentation for recursive calls + const auto new_indent = current_indent + indent_step; + if (JSON_UNLIKELY(indent_string.size() < new_indent)) + { + indent_string.resize(indent_string.size() * 2, ' '); + } + + // first n-1 elements + auto i = val.m_value.object->cbegin(); + for (std::size_t cnt = 0; cnt < val.m_value.object->size() - 1; ++cnt, ++i) + { + o->write_characters(indent_string.c_str(), new_indent); + o->write_character('\"'); + dump_escaped(i->first, ensure_ascii); + o->write_characters("\": ", 3); + dump(i->second, true, ensure_ascii, indent_step, new_indent); + o->write_characters(",\n", 2); + } + + // last element + assert(i != val.m_value.object->cend()); + assert(std::next(i) == val.m_value.object->cend()); + o->write_characters(indent_string.c_str(), new_indent); + o->write_character('\"'); + dump_escaped(i->first, ensure_ascii); + o->write_characters("\": ", 3); + dump(i->second, true, ensure_ascii, indent_step, new_indent); + + o->write_character('\n'); + o->write_characters(indent_string.c_str(), current_indent); + o->write_character('}'); + } + else + { + o->write_character('{'); + + // first n-1 elements + auto i = val.m_value.object->cbegin(); + for (std::size_t cnt = 0; cnt < val.m_value.object->size() - 1; ++cnt, ++i) + { + o->write_character('\"'); + dump_escaped(i->first, ensure_ascii); + o->write_characters("\":", 2); + dump(i->second, false, ensure_ascii, indent_step, current_indent); + o->write_character(','); + } + + // last element + assert(i != val.m_value.object->cend()); + assert(std::next(i) == val.m_value.object->cend()); + o->write_character('\"'); + dump_escaped(i->first, ensure_ascii); + o->write_characters("\":", 2); + dump(i->second, false, ensure_ascii, indent_step, current_indent); + + o->write_character('}'); + } + + return; + } + + case value_t::array: + { + if (val.m_value.array->empty()) + { + o->write_characters("[]", 2); + return; + } + + if (pretty_print) + { + o->write_characters("[\n", 2); + + // variable to hold indentation for recursive calls + const auto new_indent = current_indent + indent_step; + if (JSON_UNLIKELY(indent_string.size() < new_indent)) + { + indent_string.resize(indent_string.size() * 2, ' '); + } + + // first n-1 elements + for (auto i = val.m_value.array->cbegin(); + i != val.m_value.array->cend() - 1; ++i) + { + o->write_characters(indent_string.c_str(), new_indent); + dump(*i, true, ensure_ascii, indent_step, new_indent); + o->write_characters(",\n", 2); + } + + // last element + assert(not val.m_value.array->empty()); + o->write_characters(indent_string.c_str(), new_indent); + dump(val.m_value.array->back(), true, ensure_ascii, indent_step, new_indent); + + o->write_character('\n'); + o->write_characters(indent_string.c_str(), current_indent); + o->write_character(']'); + } + else + { + o->write_character('['); + + // first n-1 elements + for (auto i = val.m_value.array->cbegin(); + i != val.m_value.array->cend() - 1; ++i) + { + dump(*i, false, ensure_ascii, indent_step, current_indent); + o->write_character(','); + } + + // last element + assert(not val.m_value.array->empty()); + dump(val.m_value.array->back(), false, ensure_ascii, indent_step, current_indent); + + o->write_character(']'); + } + + return; + } + + case value_t::string: + { + o->write_character('\"'); + dump_escaped(*val.m_value.string, ensure_ascii); + o->write_character('\"'); + return; + } + + case value_t::boolean: + { + if (val.m_value.boolean) + { + o->write_characters("true", 4); + } + else + { + o->write_characters("false", 5); + } + return; + } + + case value_t::number_integer: + { + dump_integer(val.m_value.number_integer); + return; + } + + case value_t::number_unsigned: + { + dump_integer(val.m_value.number_unsigned); + return; + } + + case value_t::number_float: + { + dump_float(val.m_value.number_float); + return; + } + + case value_t::discarded: + { + o->write_characters("", 11); + return; + } + + case value_t::null: + { + o->write_characters("null", 4); + return; + } + } + } + + private: + /*! + @brief dump escaped string + + Escape a string by replacing certain special characters by a sequence of an + escape character (backslash) and another character and other control + characters by a sequence of "\u" followed by a four-digit hex + representation. The escaped string is written to output stream @a o. + + @param[in] s the string to escape + @param[in] ensure_ascii whether to escape non-ASCII characters with + \uXXXX sequences + + @complexity Linear in the length of string @a s. + */ + void dump_escaped(const string_t& s, const bool ensure_ascii) + { + uint32_t codepoint; + uint8_t state = UTF8_ACCEPT; + std::size_t bytes = 0; // number of bytes written to string_buffer + + for (std::size_t i = 0; i < s.size(); ++i) + { + const auto byte = static_cast(s[i]); + + switch (decode(state, codepoint, byte)) + { + case UTF8_ACCEPT: // decode found a new code point + { + switch (codepoint) + { + case 0x08: // backspace + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'b'; + break; + } + + case 0x09: // horizontal tab + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 't'; + break; + } + + case 0x0A: // newline + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'n'; + break; + } + + case 0x0C: // formfeed + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'f'; + break; + } + + case 0x0D: // carriage return + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'r'; + break; + } + + case 0x22: // quotation mark + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = '\"'; + break; + } + + case 0x5C: // reverse solidus + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = '\\'; + break; + } + + default: + { + // escape control characters (0x00..0x1F) or, if + // ensure_ascii parameter is used, non-ASCII characters + if ((codepoint <= 0x1F) or (ensure_ascii and (codepoint >= 0x7F))) + { + if (codepoint <= 0xFFFF) + { + std::snprintf(string_buffer.data() + bytes, 7, "\\u%04x", + static_cast(codepoint)); + bytes += 6; + } + else + { + std::snprintf(string_buffer.data() + bytes, 13, "\\u%04x\\u%04x", + static_cast(0xD7C0 + (codepoint >> 10)), + static_cast(0xDC00 + (codepoint & 0x3FF))); + bytes += 12; + } + } + else + { + // copy byte to buffer (all previous bytes + // been copied have in default case above) + string_buffer[bytes++] = s[i]; + } + break; + } + } + + // write buffer and reset index; there must be 13 bytes + // left, as this is the maximal number of bytes to be + // written ("\uxxxx\uxxxx\0") for one code point + if (string_buffer.size() - bytes < 13) + { + o->write_characters(string_buffer.data(), bytes); + bytes = 0; + } + break; + } + + case UTF8_REJECT: // decode found invalid UTF-8 byte + { + std::string sn(3, '\0'); + snprintf(&sn[0], sn.size(), "%.2X", byte); + JSON_THROW(type_error::create(316, "invalid UTF-8 byte at index " + std::to_string(i) + ": 0x" + sn)); + } + + default: // decode found yet incomplete multi-byte code point + { + if (not ensure_ascii) + { + // code point will not be escaped - copy byte to buffer + string_buffer[bytes++] = s[i]; + } + break; + } + } + } + + if (JSON_LIKELY(state == UTF8_ACCEPT)) + { + // write buffer + if (bytes > 0) + { + o->write_characters(string_buffer.data(), bytes); + } + } + else + { + // we finish reading, but do not accept: string was incomplete + std::string sn(3, '\0'); + snprintf(&sn[0], sn.size(), "%.2X", static_cast(s.back())); + JSON_THROW(type_error::create(316, "incomplete UTF-8 string; last byte: 0x" + sn)); + } + } + + /*! + @brief dump an integer + + Dump a given integer to output stream @a o. Works internally with + @a number_buffer. + + @param[in] x integer number (signed or unsigned) to dump + @tparam NumberType either @a number_integer_t or @a number_unsigned_t + */ + template::value or + std::is_same::value, + int> = 0> + void dump_integer(NumberType x) + { + // special case for "0" + if (x == 0) + { + o->write_character('0'); + return; + } + + const bool is_negative = (x <= 0) and (x != 0); // see issue #755 + std::size_t i = 0; + + while (x != 0) + { + // spare 1 byte for '\0' + assert(i < number_buffer.size() - 1); + + const auto digit = std::labs(static_cast(x % 10)); + number_buffer[i++] = static_cast('0' + digit); + x /= 10; + } + + if (is_negative) + { + // make sure there is capacity for the '-' + assert(i < number_buffer.size() - 2); + number_buffer[i++] = '-'; + } + + std::reverse(number_buffer.begin(), number_buffer.begin() + i); + o->write_characters(number_buffer.data(), i); + } + + /*! + @brief dump a floating-point number + + Dump a given floating-point number to output stream @a o. Works internally + with @a number_buffer. + + @param[in] x floating-point number to dump + */ + void dump_float(number_float_t x) + { + // NaN / inf + if (not std::isfinite(x)) + { + o->write_characters("null", 4); + return; + } + + // If number_float_t is an IEEE-754 single or double precision number, + // use the Grisu2 algorithm to produce short numbers which are + // guaranteed to round-trip, using strtof and strtod, resp. + // + // NB: The test below works if == . + static constexpr bool is_ieee_single_or_double + = (std::numeric_limits::is_iec559 and std::numeric_limits::digits == 24 and std::numeric_limits::max_exponent == 128) or + (std::numeric_limits::is_iec559 and std::numeric_limits::digits == 53 and std::numeric_limits::max_exponent == 1024); + + dump_float(x, std::integral_constant()); + } + + void dump_float(number_float_t x, std::true_type /*is_ieee_single_or_double*/) + { + char* begin = number_buffer.data(); + char* end = ::nlohmann::detail::to_chars(begin, begin + number_buffer.size(), x); + + o->write_characters(begin, static_cast(end - begin)); + } + + void dump_float(number_float_t x, std::false_type /*is_ieee_single_or_double*/) + { + // get number of digits for a float -> text -> float round-trip + static constexpr auto d = std::numeric_limits::max_digits10; + + // the actual conversion + std::ptrdiff_t len = snprintf(number_buffer.data(), number_buffer.size(), "%.*g", d, x); + + // negative value indicates an error + assert(len > 0); + // check if buffer was large enough + assert(static_cast(len) < number_buffer.size()); + + // erase thousands separator + if (thousands_sep != '\0') + { + const auto end = std::remove(number_buffer.begin(), + number_buffer.begin() + len, thousands_sep); + std::fill(end, number_buffer.end(), '\0'); + assert((end - number_buffer.begin()) <= len); + len = (end - number_buffer.begin()); + } + + // convert decimal point to '.' + if (decimal_point != '\0' and decimal_point != '.') + { + const auto dec_pos = std::find(number_buffer.begin(), number_buffer.end(), decimal_point); + if (dec_pos != number_buffer.end()) + { + *dec_pos = '.'; + } + } + + o->write_characters(number_buffer.data(), static_cast(len)); + + // determine if need to append ".0" + const bool value_is_int_like = + std::none_of(number_buffer.begin(), number_buffer.begin() + len + 1, + [](char c) + { + return (c == '.' or c == 'e'); + }); + + if (value_is_int_like) + { + o->write_characters(".0", 2); + } + } + + /*! + @brief check whether a string is UTF-8 encoded + + The function checks each byte of a string whether it is UTF-8 encoded. The + result of the check is stored in the @a state parameter. The function must + be called initially with state 0 (accept). State 1 means the string must + be rejected, because the current byte is not allowed. If the string is + completely processed, but the state is non-zero, the string ended + prematurely; that is, the last byte indicated more bytes should have + followed. + + @param[in,out] state the state of the decoding + @param[in,out] codep codepoint (valid only if resulting state is UTF8_ACCEPT) + @param[in] byte next byte to decode + @return new state + + @note The function has been edited: a std::array is used. + + @copyright Copyright (c) 2008-2009 Bjoern Hoehrmann + @sa http://bjoern.hoehrmann.de/utf-8/decoder/dfa/ + */ + static uint8_t decode(uint8_t& state, uint32_t& codep, const uint8_t byte) noexcept + { + static const std::array utf8d = + { + { + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 00..1F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 20..3F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 40..5F + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 60..7F + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, 9, // 80..9F + 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, 7, // A0..BF + 8, 8, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // C0..DF + 0xA, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x3, 0x4, 0x3, 0x3, // E0..EF + 0xB, 0x6, 0x6, 0x6, 0x5, 0x8, 0x8, 0x8, 0x8, 0x8, 0x8, 0x8, 0x8, 0x8, 0x8, 0x8, // F0..FF + 0x0, 0x1, 0x2, 0x3, 0x5, 0x8, 0x7, 0x1, 0x1, 0x1, 0x4, 0x6, 0x1, 0x1, 0x1, 0x1, // s0..s0 + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, // s1..s2 + 1, 2, 1, 1, 1, 1, 1, 2, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, // s3..s4 + 1, 2, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 3, 1, 1, 1, 1, 1, 1, // s5..s6 + 1, 3, 1, 1, 1, 1, 1, 3, 1, 3, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 // s7..s8 + } + }; + + const uint8_t type = utf8d[byte]; + + codep = (state != UTF8_ACCEPT) + ? (byte & 0x3fu) | (codep << 6) + : static_cast(0xff >> type) & (byte); + + state = utf8d[256u + state * 16u + type]; + return state; + } + + private: + /// the output of the serializer + output_adapter_t o = nullptr; + + /// a (hopefully) large enough character buffer + std::array number_buffer{{}}; + + /// the locale + const std::lconv* loc = nullptr; + /// the locale's thousand separator character + const char thousands_sep = '\0'; + /// the locale's decimal point character + const char decimal_point = '\0'; + + /// string buffer + std::array string_buffer{{}}; + + /// the indentation character + const char indent_char; + /// the indentation string + string_t indent_string; +}; +} +} -- cgit v1.2.3