parselink-old/include/parselink/msgpack/core/packer.h
Kurt Sassenrath 53665b8e70 Packer WIP
2023-12-08 08:46:48 -08:00

328 lines
11 KiB
C++

//-----------------------------------------------------------------------------
// ___ __ _ _
// / _ \__ _ _ __ ___ ___ / /(_)_ __ | | __
// / /_)/ _` | '__/ __|/ _ \/ / | | '_ \| |/ /
// / ___/ (_| | | \__ \ __/ /__| | | | | <
// \/ \__,_|_| |___/\___\____/_|_| |_|_|\_\ .
//
//-----------------------------------------------------------------------------
// Author: Kurt Sassenrath
// Module: msgpack
//
// Default packer implementation, which aims to deduce the best format to use
// for a given value. For example, if a 32-bit unsigned integer type only
// contains the value 5, a uint32 format would serialize into:
//
// 0xce, 0x00, 0x00, 0x00, 0x05
//
// Instead, the packer will note that this value could be stored in a positive
// fixint, which is simply:
//
// 0x05
//
// The same optimization will be applied to variable-length types, like strings,
// bytes, arrays, and maps.
//
// This flexibility comes at the cost of CPU instructions. For embedded targets,
// writer (to be renamed verbatim_packer in the future) may be a better choice.
//
// Future goals for this particular packer:
// 1. Support containers/ranges seamlessly.
// 2. Support packing of trivial POD structures without an explicit
// pack_adapter.
//
// Copyright (c) 2023 Kurt Sassenrath.
//
// License TBD.
//-----------------------------------------------------------------------------
#ifndef msgpack_core_packer_1d5939e9c1498568
#define msgpack_core_packer_1d5939e9c1498568
#include "parselink/msgpack/core/error.h"
#include "parselink/msgpack/core/format.h"
#include "parselink/msgpack/util/endianness.h"
#include <tl/expected.hpp>
#include <limits>
#include <type_traits>
namespace msgpack {
namespace detail {
// This is a generic helper function for writing integral bytes.
template <typename T, typename Itr>
constexpr auto write_integral(T value, Itr out, std::size_t sz) noexcept {
auto bytes = ::detail::as_bytes(host_to_be(value));
for (std::size_t i = 0; i < sz; ++i) {
*out++ = *(bytes.end() - sz + i);
}
return out;
}
// Depending on the format, a number of bytes will be necessary to represent
// either the value (integer formats) or the length (variable length formats).
template <format::type T>
struct pack_helper {};
template <>
struct pack_helper<format::type::unsigned_int> {
static constexpr std::size_t num_bytes(std::uint64_t value) noexcept {
#ifdef PARSELINK_CFG_MSGPACK_PACKER_USE_BRANCHES
if (count <= 0x7f) return 0;
if (count < 0x100) return 1;
if (count < 0x10000) return 2;
if (count < 0x100000000) return 4;
return 8;
#else
if (value <= std::uint64_t(format::positive_fixint::mask)) return 0;
return std::bit_ceil(std::uint64_t((std::bit_width(value) + 7) >> 3));
#endif
}
static constexpr std::byte marker(std::uint64_t value) noexcept {
switch (num_bytes(value)) {
case 0: return static_cast<std::byte>(value);
case 1: return format::uint8::marker;
case 2: return format::uint16::marker;
case 4: return format::uint32::marker;
default: return format::uint64::marker;
}
}
};
template <>
struct pack_helper<format::type::signed_int> {
static constexpr std::size_t num_bytes(std::int64_t value) noexcept {
// Probably a better way to do this.
if (value < 0 && value >= -32) return 0;
auto underlying = static_cast<std::uint64_t>(value);
// save a branch; these should be cheap on modern hardware.
std::uint64_t counts[2] = {
static_cast<std::uint64_t>(std::countl_zero(underlying)),
static_cast<std::uint64_t>(std::countl_one(underlying))};
std::uint64_t width = 1 + std::numeric_limits<std::uint64_t>::digits
- counts[underlying >> 63];
return std::bit_ceil((width + 7) >> 3);
}
static constexpr std::byte marker(std::int64_t value) noexcept {
switch (num_bytes(value)) {
case 0: return static_cast<std::byte>(value);
case 1: return format::int8::marker;
case 2: return format::int16::marker;
case 4: return format::int32::marker;
default: return format::int64::marker;
}
}
};
template <>
struct pack_helper<format::type::binary> {
static constexpr uint32_t max_supported_payload = 4;
static constexpr std::size_t num_bytes(
std::span<std::byte const> value) noexcept {
return std::min(max_supported_payload,
std::bit_ceil(std::uint32_t(
((std::bit_width(value.size()) + 7) >> 3))));
}
static constexpr auto marker(std::span<std::byte const> value) noexcept {
auto const len = num_bytes(value);
switch (len) {
case 1: return format::bin8::marker;
case 2: return format::bin16::marker;
case 4:
default: return format::bin32::marker;
}
}
};
template <>
struct pack_helper<format::type::array> {
static constexpr uint32_t max_supported_payload = 4;
static constexpr std::size_t num_bytes(std::size_t count) noexcept {
if (count < 16) return 0;
return std::min(max_supported_payload,
std::bit_ceil(std::uint32_t(
((std::bit_width(count) + 15) >> 4) << 1)));
}
static constexpr auto marker(std::size_t value) noexcept {
auto const len = num_bytes(value);
switch (len) {
case 0:
case 2: return format::array16::marker;
case 4:
default: return format::bin32::marker;
}
}
};
constexpr auto x = pack_helper<format::type::array>::num_bytes(15);
constexpr auto y = pack_helper<format::type::array>::num_bytes(16);
constexpr auto z = pack_helper<format::type::array>::num_bytes(256);
constexpr auto a = pack_helper<format::type::array>::num_bytes(65536);
template <>
struct pack_helper<format::type::string> {
static constexpr uint32_t max_supported_payload = 4;
static constexpr std::size_t num_bytes(std::string_view value) noexcept {
auto const len = value.size();
if (len <= std::uint32_t(format::fixstr::mask)) return 0;
return std::min(max_supported_payload,
std::bit_ceil(std::uint32_t((std::bit_width(len) + 7) >> 3)));
}
static constexpr std::size_t payload_size(std::string_view value) noexcept {
return value.size();
}
static constexpr auto marker(std::string_view value) noexcept {
auto const len = num_bytes(value);
switch (len) {
case 0: return format::fixstr::marker | std::byte(value.size());
case 1: return format::str8::marker;
case 2: return format::str16::marker;
case 4:
default: return format::str32::marker;
}
}
};
} // namespace detail
// Pack adapter is the basis for packing native values into MessagePack format.
template <typename T>
struct pack_adapter {};
template <typename T>
concept basic_packable_type = requires(T const& t, std::byte* b) {
{ pack_adapter<T>::total_size(t) } -> std::same_as<std::size_t>;
{ pack_adapter<T>::write(t, b) } -> std::same_as<decltype(b)>;
{ pack_adapter<T>::marker(t) } -> std::same_as<std::byte>;
};
template <typename T>
concept packable_type = basic_packable_type<T>;
template <format::type F>
struct builtin_pack_adapter {
static constexpr auto format_type = F;
using pack_helper = detail::pack_helper<F>;
// Retrieves the number of bytes to store the format, aside from the marker
// byte.
static constexpr auto format_size(auto value) noexcept {
return pack_helper::num_bytes(value);
}
// Retrieves the number of bytes to store the payload, if it is
// variable-length.
static constexpr std::size_t payload_size(auto value) noexcept {
if constexpr (requires { pack_helper::payload_size(value); }) {
return pack_helper::payload_size(value);
} else {
return 0;
}
}
static constexpr auto total_size(auto value) noexcept {
return 1 + format_size(value) + payload_size(value);
}
static constexpr auto marker(auto value) noexcept {
return detail::pack_helper<F>::marker(value);
}
};
template <std::signed_integral T>
struct pack_adapter<T> : builtin_pack_adapter<format::type::signed_int> {
template <typename Itr>
static constexpr Itr write(T value, Itr out) noexcept {
return detail::write_integral(value, out, format_size(value));
}
};
template <std::unsigned_integral T>
struct pack_adapter<T> : builtin_pack_adapter<format::type::unsigned_int> {
template <typename Itr>
static constexpr Itr write(T value, Itr out) noexcept {
return detail::write_integral(value, out, format_size(value));
}
};
template <>
struct pack_adapter<std::string_view>
: builtin_pack_adapter<format::type::string> {
template <typename Itr>
static constexpr Itr write(std::string_view value, Itr out) noexcept {
out = detail::write_integral(value.size(), out, format_size(value));
auto const* beg = reinterpret_cast<std::byte const*>(&*value.begin());
std::copy(beg, beg + value.size(), out);
return out + value.size();
}
};
template <std::convertible_to<std::string_view> T>
struct pack_adapter<T> : pack_adapter<std::string_view> {};
template <>
struct pack_adapter<std::span<std::byte const>>
: builtin_pack_adapter<format::type::binary> {
template <typename Itr>
static constexpr Itr write(
std::span<std::byte const> value, Itr out) noexcept {
out = detail::write_integral(value.size(), out, format_size(value));
std::copy(value.begin(), value.end(), out);
return out + value.size();
}
};
class packer {
using BufferType = std::span<std::byte>;
public:
constexpr packer(BufferType buff)
: buff_(buff)
, curr_(std::begin(buff_))
, end_(std::end(buff_)) {}
template <packable_type T>
constexpr tl::expected<tl::monostate, error> pack(T&& v) noexcept {
using diff_type =
std::iterator_traits<decltype(curr_)>::difference_type;
diff_type const space_needed = pack_adapter<T>::total_size(v);
if (space_needed > std::distance(curr_, end_)) {
return tl::make_unexpected(error::out_of_space);
}
*curr_++ = pack_adapter<T>::marker(v);
if (space_needed > 1) {
curr_ = pack_adapter<T>::write(v, curr_);
}
return tl::monostate{};
}
constexpr auto tell() const noexcept {
return std::distance(std::begin(buff_), curr_);
}
constexpr auto subspan() const noexcept { return buff_.subspan(0, tell()); }
private:
BufferType buff_;
decltype(std::begin(buff_)) curr_;
decltype(std::end(buff_)) end_;
};
} // namespace msgpack
#endif // msgpack_core_packer_1d5939e9c1498568