amorfo77/json
1
0 Fork 0
mirror of https://github.com/nlohmann/json.git synced 2026-10-22 22:08:55 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
4cde5e39b9d255ba5205ae279e0297480a534b08
json/include/nlohmann/detail/input/json_sax.hpp
T
Niels Lohmann 4cde5e39b9 Collect the members of ordered_json objects on a stack until they are known 
An ordered_json object keeps its members in a vector and appends every new
member to it, so reading an object of unknown size reallocated the vector at
1, 2, 4, 8, ... members, and ordered_map copies all keys on every
reallocation. The DOM SAX parser now collects the members of such objects on
one stack shared by all open objects, as it does the elements of arrays, and
moves them into a vector of exactly the right size when the closing brace
arrives. An object that gets past 32 members switches to appending to its
own vector (and, from 64, to the key index of the earlier commit). The
members are the same, in the same order, and a key that occurs twice keeps
the place of its first occurrence. std::map based objects, objects whose
size the input announces (binary formats), and builds with JSON_DIAGNOSTICS
or JSON_DIAGNOSTIC_POSITIONS are untouched.

What only ordered objects need (the collected members, the key indexes) is
a member of the parser only for them, and the code that stores a value tries
the array that takes its elements directly first. An array that outgrows
the stack now reserves 32 elements, the capacity a vector that grew by
itself has at that point; with the larger reserve before, parsing
floats.json (1 000 000 numbers) was 2 % slower on Apple M1 than before this
series.

ordered_json parse, Linux Ir against develop (GCC / Clang): citm -11.6 % /
-11.7 %, twitter -8.4 % / -8.4 %, jeopardy (12 000 objects of 7 keys)
-10.5 % / -8.0 %, canada -3.9 % / -5.2 %, 124-byte object -20.9 % / -16.5 %.
macOS time (M1, earlier base): citm -19 %, twitter -13 %, jeopardy -13 %,
canada -14 %, 124-byte object -26 %. json parse: steady-state twitter -0.1 %
(see the previous commits), citm -1.2 %.

Signed-off-by: Niels Lohmann <mail@nlohmann.me>
2026-10-22 09:28:08 +02:00

2198 lines
74 KiB
C++
Raw Blame History

// __ _____ _____ _____
// __| | __| | | | JSON for Modern C++
// | | |__ | | | | | | version 3.12.0
// |_____|_____|_____|_|___| https://github.com/nlohmann/json
//
// SPDX-FileCopyrightText: 2013-2026 Niels Lohmann <https://nlohmann.me>
// SPDX-License-Identifier: MIT
#pragma once
#include <algorithm> // copy, find_if, min
#include <cstddef>
#include <functional> // hash
#include <limits> // numeric_limits
#include <memory> // unique_ptr
#include <new> // operator new
#include <string> // string
#include <type_traits> // enable_if_t
#include <utility> // move, pair
#include <vector> // vector
#include <nlohmann/detail/exceptions.hpp>
#include <nlohmann/detail/input/lexer.hpp>
#include <nlohmann/detail/json_custom_base_class.hpp>
#include <nlohmann/detail/macro_scope.hpp>
#include <nlohmann/detail/meta/cpp_future.hpp>
#include <nlohmann/detail/meta/detected.hpp>
#include <nlohmann/detail/string_concat.hpp>
NLOHMANN_JSON_NAMESPACE_BEGIN
/*!
@brief SAX interface
This class describes the SAX interface used by @ref nlohmann::json::sax_parse.
Each function is called in different situations while the input is parsed. The
boolean return value informs the parser whether to continue processing the
input.
*/
template<typename BasicJsonType>
struct json_sax
{
using number_integer_t = typename BasicJsonType::number_integer_t;
using number_unsigned_t = typename BasicJsonType::number_unsigned_t;
using number_float_t = typename BasicJsonType::number_float_t;
using string_t = typename BasicJsonType::string_t;
using binary_t = typename BasicJsonType::binary_t;
/*!
@brief a null value was read
@return whether parsing should proceed
*/
virtual bool null() = 0;
/*!
@brief a boolean value was read
@param[in] val boolean value
@return whether parsing should proceed
*/
virtual bool boolean(bool val) = 0;
/*!
@brief an integer number was read
@param[in] val integer value
@return whether parsing should proceed
*/
virtual bool number_integer(number_integer_t val) = 0;
/*!
@brief an unsigned integer number was read
@param[in] val unsigned integer value
@return whether parsing should proceed
*/
virtual bool number_unsigned(number_unsigned_t val) = 0;
/*!
@brief a floating-point number was read
@param[in] val floating-point value
@param[in] s raw token value
@return whether parsing should proceed
*/
virtual bool number_float(number_float_t val, const string_t& s) = 0;
/*!
@brief a string value was read
@param[in] val string value
@return whether parsing should proceed
@note It is safe to move the passed string value.
*/
virtual bool string(string_t& val) = 0;
/*!
@brief a binary value was read
@param[in] val binary value
@return whether parsing should proceed
@note It is safe to move the passed binary value.
*/
virtual bool binary(binary_t& val) = 0;
/*!
@brief the beginning of an object was read
@param[in] elements number of object elements or -1 if unknown
@return whether parsing should proceed
@note binary formats may report the number of elements
*/
virtual bool start_object(std::size_t elements) = 0;
/*!
@brief an object key was read
@param[in] val object key
@return whether parsing should proceed
@note It is safe to move the passed string.
*/
virtual bool key(string_t& val) = 0;
/*!
@brief the end of an object was read
@return whether parsing should proceed
*/
virtual bool end_object() = 0;
/*!
@brief the beginning of an array was read
@param[in] elements number of array elements or -1 if unknown
@return whether parsing should proceed
@note binary formats may report the number of elements
*/
virtual bool start_array(std::size_t elements) = 0;
/*!
@brief the end of an array was read
@return whether parsing should proceed
*/
virtual bool end_array() = 0;
/*!
@brief a parse error occurred
@param[in] position the position in the input where the error occurs
@param[in] last_token the last read token
@param[in] ex an exception object describing the error
@return whether parsing should proceed (must return false)
*/
virtual bool parse_error(std::size_t position,
const std::string& last_token,
const detail::exception& ex) = 0;
json_sax() = default;
json_sax(const json_sax&) = default;
json_sax(json_sax&&) noexcept = default;
json_sax& operator=(const json_sax&) = default;
json_sax& operator=(json_sax&&) noexcept = default;
virtual ~json_sax() = default;
};
namespace detail
{
constexpr std::size_t unknown_size()
{
return (std::numeric_limits<std::size_t>::max)();
}
/*!
@brief reserve capacity for @a len elements in array @a arr
Reserving upfront avoids repeated reallocations while the elements are added,
but the reservation is capped so a bogus/hostile length (which is not bounded
by max_size(), unlike e.g. std::vector) cannot trigger an oversized allocation
for a small or truncated input.
The overload below is selected for array types without reserve() (e.g.,
std::deque), which are then left untouched.
*/
template<typename ArrayType>
auto reserve_array(ArrayType& arr, std::size_t len, priority_tag<1> /*unused*/)
-> decltype(arr.reserve(len), void())
{
constexpr std::size_t reserve_cap = 16384;
arr.reserve((std::min)(len, reserve_cap));
}
template<typename ArrayType>
inline void reserve_array(ArrayType& /*arr*/, std::size_t /*len*/, priority_tag<0> /*unused*/)
{}
/*!
@brief reports an object or array whose announced size exceeds max_size()
Shared by json_sax_dom_parser and json_sax_dom_callback_parser.
@param[in] sax the SAX parser to report the error to
@param[in] len the number of elements announced by the input, or unknown_size()
@param[in] kind "object" or "array"
@param[in] ref the object or array that was just created
@return whether parsing should continue (false after reporting out_of_range.408)
*/
template<typename SAX, typename BasicJsonType>
JSON_HEDLEY_NEVER_INLINE
bool container_too_large_error(SAX& sax, std::size_t len, const char* kind, BasicJsonType* ref)
{
return sax.parse_error(0, "", out_of_range::create(exception_id::container_too_large, concat("excessive ", kind, " size: ", std::to_string(len)), ref));
}
template<typename SAX, typename BasicJsonType>
bool check_container_size(SAX& sax, std::size_t len, const char* kind, BasicJsonType* ref)
{
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > ref->max_size()))
{
return container_too_large_error(sax, len, kind, ref);
}
return true;
}
#if JSON_DIAGNOSTIC_POSITIONS
/*!
@brief set the diagnostic positions of a value the DOM SAX parsers just stored
Shared by json_sax_dom_parser and json_sax_dom_callback_parser. basic_json
befriends this struct, as the position members are private.
*/
struct diagnostic_positions
{
/*!
@param[in,out] v the object or array whose opening brace or bracket was just read
@param[in] lexer the lexer that read it, or nullptr to leave @a v alone
*/
template<typename BasicJsonType, typename LexerType>
static void set_container_start(BasicJsonType& v, LexerType* lexer)
{
if (lexer)
{
// Lexer has read the first character of the container, so
// subtract 1 from the position to get the correct start position.
v.start_position = lexer->get_chars_read_total() - 1;
}
}
/*!
@param[in,out] v the object or array whose closing brace or bracket was just read
@param[in] lexer the lexer that read it, or nullptr to leave @a v alone
*/
template<typename BasicJsonType, typename LexerType>
static void set_container_end(BasicJsonType& v, LexerType* lexer)
{
if (lexer)
{
// Lexer's position is past the closing brace or bracket, so set that as the end position.
v.end_position = lexer->get_chars_read_total();
}
}
/*!
@param[in,out] v the value that was just parsed
@param[in] lexer the lexer that read it, or nullptr to leave @a v alone
*/
template<typename BasicJsonType, typename LexerType>
static void set_from_lexer(BasicJsonType& v, LexerType* lexer)
{
if (lexer)
{
// Lexer has read past the current field value, so set the end position to the current position.
// The start position will be set below based on the length of the string representation
// of the value.
v.end_position = lexer->get_chars_read_total();
switch (v.type())
{
case value_t::boolean:
{
// 4 and 5 are the string length of "true" and "false"
v.start_position = v.end_position - (v.m_data.m_value.boolean ? 4 : 5);
break;
}
case value_t::null:
{
// 4 is the string length of "null"
v.start_position = v.end_position - 4;
break;
}
case value_t::string:
{
// escape sequences make the token longer than the value it
// parses to, so the start position cannot be derived from
// the value; use the offset the lexer recorded instead
v.start_position = lexer->get_token_start_position();
break;
}
case value_t::discarded:
{
// an object or array the callback of
// json_sax_dom_callback_parser rejected has no position
v.end_position = std::string::npos;
v.start_position = v.end_position;
break;
}
case value_t::binary:
case value_t::number_integer:
case value_t::number_unsigned:
case value_t::number_float:
{
v.start_position = v.end_position - lexer->get_string().size();
break;
}
case value_t::object:
case value_t::array:
{
// object and array are handled in start_object() and start_array() handlers
// skip setting the values here.
break;
}
default: // LCOV_EXCL_LINE
// Handle all possible types discretely, default handler should never be reached.
JSON_ASSERT(false); // NOLINT(cert-dcl03-c,hicpp-static-assert,misc-static-assert) LCOV_EXCL_LINE
}
}
}
};
#endif
/*!
@brief a hash table that finds the position of a key in an object whose
elements are kept in a vector (ordered_map)
The elements stay where they are: the table only holds their hashes and
positions, and the caller compares the keys.
*/
class object_key_index
{
public:
static constexpr std::size_t npos = static_cast<std::size_t>(-1);
/// @param[in] element_hashes the hashes of the keys of the elements that are there already
explicit object_key_index(std::vector<std::size_t>&& element_hashes)
: hashes(std::move(element_hashes))
{
std::size_t size = 256;
while (size < 4 * hashes.size())
{
size *= 2;
}
rebuild(size);
}
/// @return the position of the element with @a hash for which @a equal(position) holds, or npos
template<typename Equal>
std::size_t find(const std::size_t hash, Equal equal) const
{
for (std::size_t slot = hash & mask;; slot = (slot + 1) & mask)
{
const std::size_t entry = table[slot];
if (entry == 0)
{
return npos;
}
if (hashes[entry - 1] == hash && equal(entry - 1))
{
return entry - 1;
}
}
}
/// add the next element, whose position is the number of elements so far
void add(const std::size_t hash)
{
hashes.push_back(hash);
if (hashes.size() * 2 > table.size())
{
rebuild(table.size() * 2);
}
else
{
place(hash, hashes.size());
}
}
private:
void rebuild(const std::size_t size)
{
table.assign(size, 0);
mask = size - 1;
for (std::size_t i = 0; i < hashes.size(); ++i)
{
place(hashes[i], i + 1);
}
}
void place(const std::size_t hash, const std::size_t entry)
{
std::size_t slot = hash & mask;
while (table[slot] != 0)
{
slot = (slot + 1) & mask;
}
table[slot] = entry;
}
/// the hash of each element, by position
std::vector<std::size_t> hashes;
/// open addressing, a power of two in size: position + 1 of an element, or 0 for a free slot
std::vector<std::size_t> table {};
std::size_t mask = 0;
};
/*!
@brief SAX implementation to create a JSON value from SAX events
This class implements the @ref json_sax interface and processes the SAX events
to create a JSON value which makes it basically a DOM parser. The structure or
hierarchy of the JSON value is managed by the stack `frames` which contains
an entry for each open array or object. The parser of the text formats, for
which the size of a container is not known up front, collects the elements of
an array and the members of an `ordered_json` object on a stack and moves them
into a container of the exact size; the binary readers use the plain
@ref json_sax_dom_binary_parser instead.
After successful parsing, the value that is passed by reference to the
constructor contains the parsed value.
@tparam BasicJsonType the JSON type
*/
template<typename BasicJsonType, typename InputAdapterType>
class json_sax_dom_parser
{
public:
using number_integer_t = typename BasicJsonType::number_integer_t;
using number_unsigned_t = typename BasicJsonType::number_unsigned_t;
using number_float_t = typename BasicJsonType::number_float_t;
using string_t = typename BasicJsonType::string_t;
using binary_t = typename BasicJsonType::binary_t;
using lexer_t = lexer<BasicJsonType, InputAdapterType>;
/*!
@param[in,out] r reference to a JSON value that is manipulated while
parsing
@param[in] allow_exceptions_ whether parse errors yield exceptions
*/
explicit json_sax_dom_parser(BasicJsonType& r, const bool allow_exceptions_ = true, lexer_t* lexer_ = nullptr)
: root(r), allow_exceptions(allow_exceptions_), m_lexer_ref(lexer_)
{}
// make class move-only
json_sax_dom_parser(const json_sax_dom_parser&) = delete;
json_sax_dom_parser(json_sax_dom_parser&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor)
json_sax_dom_parser& operator=(const json_sax_dom_parser&) = delete;
json_sax_dom_parser& operator=(json_sax_dom_parser&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor)
~json_sax_dom_parser() = default;
bool null()
{
handle_value(nullptr);
return true;
}
bool boolean(bool val)
{
handle_value(val);
return true;
}
bool number_integer(number_integer_t val)
{
handle_value(val);
return true;
}
bool number_unsigned(number_unsigned_t val)
{
handle_value(val);
return true;
}
bool number_float(number_float_t val, const string_t& /*unused*/)
{
handle_value(val);
return true;
}
bool string(string_t& val)
{
// json_sax documents that the passed value may be moved from,
// so hand the buffer over instead of copying it
handle_value(std::move(val));
return true;
}
bool binary(binary_t& val)
{
// json_sax documents that the passed value may be moved from,
// so hand the buffer over instead of copying it
handle_value(std::move(val));
return true;
}
bool start_object(std::size_t len)
{
#if JSON_DIAGNOSTICS || JSON_DIAGNOSTIC_POSITIONS
// the values need their final place from the start (their parent
// pointers and positions refer to it), so no buffering
return start_stored_object(len);
#else
return start_object(len, std::integral_constant<bool, indexed_keys> {});
#endif
}
bool key(string_t& val)
{
JSON_ASSERT(!frames.empty());
return key(val, std::integral_constant<bool, indexed_keys> {});
}
bool end_object()
{
JSON_ASSERT(!frames.empty());
return end_object(std::integral_constant<bool, indexed_keys> {});
}
bool start_array(std::size_t len)
{
if (JSON_HEDLEY_UNLIKELY(len != detail::unknown_size() && len > array_t().max_size()))
{
// the array does not exist yet: report with an empty one
BasicJsonType too_large(BasicJsonType::value_t::array);
return container_too_large_error(*this, len, "array", &too_large);
}
#if JSON_DIAGNOSTICS || JSON_DIAGNOSTIC_POSITIONS
// the values need their final place from the start (their parent
// pointers and positions refer to it), so no buffering
start_stored_array(len);
#else
if (len != detail::unknown_size())
{
// the number of elements is known: reserve and store them directly
start_stored_array(len);
}
else
{
// the number of elements is not known, so growing the array
// would reallocate it (and, being a vector, move its elements)
// again and again. Collect the elements on a stack shared by all
// arrays instead, and copy them to an array of the exact size
// when the closing bracket arrives.
BasicJsonType* slot = nullptr;
std::size_t dest_member = 0;
container_destination(slot, dest_member);
reserve_elements();
frames.push_back(frame{nullptr, nullptr, slot, elements.size(), ordered_state(dest_member, false)});
in_object = false;
stored_array = nullptr;
}
#endif
return true;
}
bool end_array()
{
JSON_ASSERT(!frames.empty());
JSON_ASSERT(frames.back().object == nullptr);
const frame f = frames.back();
frames.pop_back();
enter_enclosing_container();
if (f.array == nullptr)
{
// the elements are on the stack: make them the array
const auto first = elements.begin() + static_cast<std::ptrdiff_t>(f.start);
array_t* const arr = BasicJsonType::template create<array_t>(std::make_move_iterator(first), std::make_move_iterator(elements.end()));
BasicJsonType value;
value.m_data.m_type = BasicJsonType::value_t::array;
value.m_data.m_value.array = arr;
elements.erase(first, elements.end());
place_container(f, frames.empty() ? nullptr : &frames.back(), std::move(value), true);
return true;
}
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_container_end(*f.slot, m_lexer_ref);
#endif
#if JSON_DIAGNOSTICS
f.slot->set_parents();
#endif
return true;
}
template<class Exception>
bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/,
const Exception& ex)
{
errored = true;
static_cast<void>(ex);
if (allow_exceptions)
{
JSON_THROW(ex);
}
return false;
}
constexpr bool is_errored() const
{
return errored;
}
private:
using object_t = typename BasicJsonType::object_t;
using array_t = typename BasicJsonType::array_t;
/// whether objects keep their elements in a vector and a key can be hashed,
/// so that a big object can find the keys it has in a hash table instead of
/// searching all of them
template<typename ObjectType, typename StringType>
using detect_indexable_keys = decltype(std::declval<ObjectType&>().emplace_new_key(std::declval<const StringType&>(), std::declval<typename ObjectType::mapped_type>()),
std::hash<StringType>()(std::declval<const StringType&>()));
static constexpr bool indexed_keys = is_detected<detect_indexable_keys, object_t, string_t>::value;
/// what a frame holds for objects that are kept in a vector: the key index
/// of an object with many keys, and for an object that is collected on the
/// stack the member of the enclosing object it will become
struct ordered_frame
{
/// the entry of key_indexes of an object that has one, else nullptr
object_key_index* index;
/// object or array without a slot in an object collected on the stack:
/// the index of the member in `members` it goes to
std::size_t dest_member;
/// whether this is an object collected on the stack
bool buffered;
};
struct no_ordered_frame {};
using ordered_frame_state = typename std::conditional<indexed_keys, ordered_frame, no_ordered_frame>::type;
static ordered_frame make_ordered_state(const std::size_t dest_member, const bool buffered, std::true_type /*indexed*/) noexcept
{
ordered_frame state = {nullptr, dest_member, buffered};
return state;
}
static no_ordered_frame make_ordered_state(const std::size_t /*dest_member*/, const bool /*buffered*/, std::false_type /*indexed*/) noexcept
{
return no_ordered_frame();
}
static ordered_frame_state ordered_state(const std::size_t dest_member, const bool buffered) noexcept
{
return make_ordered_state(dest_member, buffered, std::integral_constant<bool, indexed_keys> {});
}
/// an array or object that is being read
struct frame
{
/// the object, or nullptr for an array
object_t* object;
/// the array if its elements are stored in it as they arrive, or
/// nullptr for an array and an object
array_t* array;
/// array without `array`: where the finished array goes (nullptr: it
/// is an element of the enclosing array); otherwise: the value that
/// holds the object or array
BasicJsonType* slot;
/// array without `array`: the index of its first element in `elements`
std::size_t start;
/// what objects in a vector add (see ordered_frame)
ordered_frame_state ordered;
};
/// the open arrays and objects; the first few live in the object itself, so
/// that the common documents do not allocate for them
class frame_stack
{
public:
frame_stack() = default;
frame_stack(const frame_stack&) = delete;
frame_stack& operator=(const frame_stack&) = delete;
frame_stack(frame_stack&& other) noexcept
{
take(other);
}
frame_stack& operator=(frame_stack&& other) noexcept
{
if (this != &other)
{
release();
take(other);
}
return *this;
}
~frame_stack()
{
release();
}
bool empty() const noexcept
{
return count == 0;
}
std::size_t size() const noexcept
{
return count;
}
frame& back() noexcept
{
return data[count - 1];
}
frame& operator[](const std::size_t i) noexcept
{
return data[i];
}
void push_back(const frame& f)
{
if (JSON_HEDLEY_UNLIKELY(count == capacity))
{
grow();
}
data[count++] = f;
}
void pop_back() noexcept
{
JSON_ASSERT(count > 0);
--count;
}
private:
JSON_HEDLEY_NEVER_INLINE
void grow()
{
const std::size_t new_capacity = capacity * 2;
frame* const new_data = new frame[new_capacity];
std::copy(data, data + count, new_data);
release();
data = new_data;
capacity = new_capacity;
}
void release() noexcept
{
if (data != local)
{
delete[] data;
data = local;
capacity = local_capacity;
}
}
/// make the contents of @a other our own
void take(frame_stack& other) noexcept
{
if (other.data == other.local)
{
std::copy(other.local, other.local + other.count, local);
data = local;
capacity = local_capacity;
}
else
{
data = other.data;
capacity = other.capacity;
other.data = other.local;
other.capacity = local_capacity;
}
count = other.count;
other.count = 0;
}
static constexpr std::size_t local_capacity = 8;
frame local[local_capacity];
frame* data = local;
std::size_t count = 0;
std::size_t capacity = local_capacity;
};
void release_key_index(const frame& f, std::true_type /*indexed*/)
{
if (JSON_HEDLEY_UNLIKELY(f.ordered.index != nullptr))
{
ordered_data.key_indexes.pop_back();
}
}
static void release_key_index(const frame& /*f*/, std::false_type /*indexed*/) noexcept {}
/// the number of keys from which an object looks them up in a hash table
static constexpr std::size_t key_index_threshold = 64;
/// add @a key to the object of @a f (or find it, if it is there), searching it in std::map
static BasicJsonType* insert_key(frame& f, string_t&& key, std::false_type /*indexed*/)
{
return &f.object->operator[](std::move(key));
}
/// add @a key to the object of @a f (or find it, if it is there), searching it in a vector
BasicJsonType* insert_key(frame& f, string_t&& key, std::true_type /*indexed*/)
{
object_t& obj = *f.object;
if (JSON_HEDLEY_LIKELY(obj.size() < key_index_threshold))
{
return &obj[std::move(key)];
}
return insert_key_with_index(f, std::move(key));
}
/// add @a key to the big object of @a f (or find it, if it is there), using its key index
JSON_HEDLEY_NEVER_INLINE
BasicJsonType* insert_key_with_index(frame& f, string_t&& key)
{
object_t& obj = *f.object;
const std::hash<string_t> hasher {};
const std::size_t hash = hasher(key);
if (JSON_HEDLEY_UNLIKELY(f.ordered.index == nullptr))
{
std::vector<std::size_t> hashes;
hashes.reserve(obj.size());
for (const auto& element : obj)
{
hashes.push_back(hasher(element.first));
}
ordered_data.key_indexes.emplace_back(new object_key_index(std::move(hashes)));
f.ordered.index = ordered_data.key_indexes.back().get();
}
const std::size_t position = f.ordered.index->find(hash, [&obj, &key](std::size_t i)
{
return (obj.begin() + static_cast<std::ptrdiff_t>(i))->first == key;
});
if (position != object_key_index::npos)
{
return &(obj.begin() + static_cast<std::ptrdiff_t>(position))->second;
}
obj.emplace_new_key(key, BasicJsonType());
f.ordered.index->add(hash);
return &obj.back().second;
}
/// where an array or object that starts now will go when it is finished:
/// @a slot is the value to assign it to (the root or a member of an object
/// that is stored directly), or nullptr if it is an element of the
/// enclosing array or a member of the enclosing object that is collected on
/// the stack, in which case @a dest_member is the member
void container_destination(BasicJsonType*& slot, std::size_t& dest_member)
{
if (frames.empty())
{
slot = &root;
}
else if (frames.back().object != nullptr)
{
slot = object_element;
}
else
{
collected_object_member(frames.back(), dest_member, std::integral_constant<bool, indexed_keys> {});
}
}
void collected_object_member(const frame& enclosing, std::size_t& dest_member, std::true_type /*indexed*/) const noexcept
{
if (enclosing.ordered.buffered)
{
dest_member = ordered_data.member_index;
}
}
static void collected_object_member(const frame& /*enclosing*/, std::size_t& /*dest_member*/, std::false_type /*indexed*/) noexcept {}
static bool is_collected_object(const frame& f, std::true_type /*indexed*/) noexcept
{
return f.ordered.buffered;
}
static bool is_collected_object(const frame& /*f*/, std::false_type /*indexed*/) noexcept
{
return false;
}
/// put a finished array or object that was collected on the stack, or
/// that outgrew it, where it belongs; @a enclosing is the container it is in.
/// @a enclosing_is_innermost tells whether that is the container that gets
/// the next value, so that it can be switched to storing its elements
/// directly if the value was the one too many.
void place_container(const frame& f, frame* enclosing, BasicJsonType&& value, const bool enclosing_is_innermost)
{
if (f.slot != nullptr)
{
*f.slot = std::move(value);
}
else if (is_collected_object(*enclosing, std::integral_constant<bool, indexed_keys> {}))
{
place_member(f, std::move(value), std::integral_constant<bool, indexed_keys> {});
}
else if (enclosing->array != nullptr)
{
enclosing->array->emplace_back(std::move(value));
}
else
{
elements.emplace_back(std::move(value));
if (JSON_HEDLEY_UNLIKELY(enclosing_is_innermost && elements.size() - enclosing->start > buffer_limit))
{
spill(*enclosing);
}
}
}
void place_member(const frame& f, BasicJsonType&& value, std::true_type /*indexed*/)
{
ordered_data.members[f.ordered.dest_member].second = std::move(value);
}
static void place_member(const frame& /*f*/, BasicJsonType&& /*value*/, std::false_type /*indexed*/) noexcept {}
/// store the members of a new object directly in it
bool start_stored_object(std::size_t len)
{
BasicJsonType* const ref = store(BasicJsonType::value_t::object);
frames.push_back(frame{ref->m_data.m_value.object, nullptr, ref, 0, ordered_state(0, false)});
in_object = true;
stored_array = nullptr;
#if JSON_DIAGNOSTIC_POSITIONS
// Ensure this is after the call to store to ensure correct start position.
diagnostic_positions::set_container_start(*ref, m_lexer_ref);
#endif
return check_container_size(*this, len, "object", ref);
}
/// objects in std::map and the like have no use for collecting their members
bool start_object(std::size_t len, std::false_type /*indexed*/)
{
return start_stored_object(len);
}
/// An object in a vector of members adds a member by appending it to the
/// vector, which grows it again and again (and copies the keys each time).
/// Collect the members of an object of unknown size on a stack shared by
/// all open objects instead, and move them into a vector of the exact size
/// when the closing brace arrives.
bool start_object(std::size_t len, std::true_type /*indexed*/)
{
if (len != detail::unknown_size())
{
return start_stored_object(len);
}
BasicJsonType* slot = nullptr;
std::size_t dest_member = 0;
container_destination(slot, dest_member);
frames.push_back(frame{nullptr, nullptr, slot, ordered_data.members.size(), ordered_state(dest_member, true)});
in_object = false;
stored_array = nullptr;
return true;
}
bool key(string_t& val, std::false_type /*indexed*/)
{
JSON_ASSERT(frames.back().object != nullptr);
// add null at the given key and store the reference for later; the
// key is moved, as the lexer starts the next token with reset()
object_element = insert_key(frames.back(), std::move(val), std::false_type {});
return true;
}
bool key(string_t& val, std::true_type /*indexed*/)
{
frame& f = frames.back();
if (JSON_HEDLEY_UNLIKELY(f.object == nullptr))
{
JSON_ASSERT(f.ordered.buffered);
if (collect_member(f, val))
{
return true;
}
// the object has grown big: it is stored directly from here on
}
object_element = insert_key(f, std::move(val), std::true_type {});
return true;
}
/*!
@brief add a member to the object that is collected on the stack, or find the one with the key
@return false if the object has collected enough members to be stored
directly, which is done and which leaves the key to be inserted
*/
bool collect_member(frame& f, string_t& key)
{
for (std::size_t i = f.start; i < ordered_data.members.size(); ++i)
{
if (ordered_data.members[i].first == key)
{
ordered_data.member_index = i;
return true;
}
}
if (JSON_HEDLEY_UNLIKELY(ordered_data.members.size() - f.start >= object_buffer_limit))
{
spill_object(f);
return false;
}
reserve_members();
ordered_data.members.emplace_back(std::move(key), BasicJsonType());
ordered_data.member_index = ordered_data.members.size() - 1;
return true;
}
/// the members of an object that is collected on the stack as an object
object_t* members_to_object(const frame& f, BasicJsonType& value)
{
const auto first = ordered_data.members.begin() + static_cast<std::ptrdiff_t>(f.start);
object_t* const obj = BasicJsonType::template create<object_t>(std::make_move_iterator(first), std::make_move_iterator(ordered_data.members.end()));
value.m_data.m_type = BasicJsonType::value_t::object;
value.m_data.m_value.object = obj;
ordered_data.members.erase(first, ordered_data.members.end());
return obj;
}
/// switch an object that collects its members on the stack to storing them directly
JSON_HEDLEY_NEVER_INLINE
void spill_object(frame& f)
{
BasicJsonType value;
f.object = members_to_object(f, value);
f.ordered.buffered = false;
in_object = true;
place_container(f, frames.size() >= 2 ? &frames[frames.size() - 2] : nullptr, std::move(value), false);
}
bool end_object(std::false_type /*indexed*/)
{
return end_stored_object();
}
bool end_object(std::true_type /*indexed*/)
{
if (JSON_HEDLEY_LIKELY(frames.back().object != nullptr))
{
return end_stored_object();
}
// the members are on the stack: make them the object
const frame f = frames.back();
frames.pop_back();
enter_enclosing_container();
BasicJsonType value;
members_to_object(f, value);
place_container(f, frames.empty() ? nullptr : &frames.back(), std::move(value), true);
return true;
}
bool end_stored_object()
{
JSON_ASSERT(frames.back().object != nullptr);
#if JSON_DIAGNOSTICS || JSON_DIAGNOSTIC_POSITIONS
// objects are never collected, so their value is where it was stored
BasicJsonType* const ref = frames.back().slot;
#endif
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_container_end(*ref, m_lexer_ref);
#endif
#if JSON_DIAGNOSTICS
ref->set_parents();
#endif
release_key_index(frames.back(), std::integral_constant<bool, indexed_keys> {});
frames.pop_back();
enter_enclosing_container();
return true;
}
/// the number of members an object collects on the stack before it
/// switches to storing them directly
static constexpr std::size_t object_buffer_limit = 32;
void reserve_members()
{
if (JSON_HEDLEY_UNLIKELY(ordered_data.members.capacity() == 0))
{
ordered_data.members.reserve(16);
}
}
/// the number of elements an array collects on the stack before it
/// switches to storing them directly, so that it does not hold up the
/// stack with its elements (and move all of them in the end)
static constexpr std::size_t buffer_limit = 16;
void reserve_elements()
{
if (JSON_HEDLEY_UNLIKELY(elements.capacity() == 0))
{
elements.reserve(32);
}
}
/// store the elements of a new array directly in it
void start_stored_array(const std::size_t len)
{
BasicJsonType* const ref = store(BasicJsonType::value_t::array);
frames.push_back(frame{nullptr, ref->m_data.m_value.array, ref, 0, ordered_state(0, false)});
in_object = false;
stored_array = ref->m_data.m_value.array;
#if JSON_DIAGNOSTIC_POSITIONS
// Ensure this is after the call to store to ensure correct start position.
diagnostic_positions::set_container_start(*ref, m_lexer_ref);
#endif
if (len != detail::unknown_size())
{
reserve_array(*ref->m_data.m_value.array, len, priority_tag<1> {});
}
}
/*!
@brief store a value as the next element of the open array or object (or
as the root)
@return the stored value
*/
template<typename Value>
JSON_HEDLEY_RETURNS_NON_NULL
BasicJsonType* store(Value&& v)
{
if (stored_array != nullptr)
{
stored_array->emplace_back(std::forward<Value>(v));
return &stored_array->back();
}
if (in_object)
{
JSON_ASSERT(object_element);
set_member(*object_element, std::forward<Value>(v), std::is_same<typename BasicJsonType::json_base_class_t, json_default_base> {});
return object_element;
}
return store_collected(std::forward<Value>(v), std::integral_constant<bool, indexed_keys> {});
}
/// store a value that is the root, an element of an array that is collected
/// on the stack, or (in a vector) a member of an object that is
template<typename Value>
BasicJsonType* store_collected(Value&& v, std::true_type /*indexed*/)
{
if (!frames.empty() && frames.back().ordered.buffered)
{
BasicJsonType& member = ordered_data.members[ordered_data.member_index].second;
set_member(member, std::forward<Value>(v), std::is_same<typename BasicJsonType::json_base_class_t, json_default_base> {});
return &member;
}
return store_collected(std::forward<Value>(v), std::false_type {});
}
template<typename Value>
BasicJsonType* store_collected(Value&& v, std::false_type /*indexed*/)
{
if (frames.empty())
{
root = BasicJsonType(std::forward<Value>(v));
return &root;
}
frame& f = frames.back();
elements.emplace_back(std::forward<Value>(v));
if (JSON_HEDLEY_UNLIKELY(elements.size() - f.start > buffer_limit))
{
return spill(f);
}
return &elements.back();
}
/*!
@brief switch an array that collects its elements on the stack to storing
them directly
@return the last element of the array
*/
JSON_HEDLEY_NEVER_INLINE
BasicJsonType* spill(frame& f)
{
JSON_ASSERT(f.array == nullptr);
const auto first = elements.begin() + static_cast<std::ptrdiff_t>(f.start);
BasicJsonType value(BasicJsonType::value_t::array);
array_t& arr = *value.m_data.m_value.array;
// the capacity a vector that grew by itself has with this many elements
reserve_array(arr, 2 * buffer_limit, priority_tag<1> {});
arr.insert(arr.end(), std::make_move_iterator(first), std::make_move_iterator(elements.end()));
elements.erase(first, elements.end());
f.array = &arr;
stored_array = &arr;
// the array takes the place its first element had on the stack
place_container(f, frames.size() >= 2 ? &frames[frames.size() - 2] : nullptr, std::move(value), false);
return &arr.back();
}
/// set the value of the member key() just added
template<typename Value>
JSON_HEDLEY_ALWAYS_INLINE
static void set_member(BasicJsonType& member, Value&& v, std::true_type /*default base*/)
{
if (JSON_HEDLEY_LIKELY(member.is_null()))
{
// a new member is null and holds nothing: construct the value
// over it instead of destroying and swapping
::new (static_cast<void*>(&member)) BasicJsonType(std::forward<Value>(v));
}
else
{
// the key occurred before
member = BasicJsonType(std::forward<Value>(v));
}
}
/// a custom base class may care about the lifetime of the value
template<typename Value>
static void set_member(BasicJsonType& member, Value&& v, std::false_type /*custom base*/)
{
member = BasicJsonType(std::forward<Value>(v));
}
/// after a container was closed: the one it was in is the one that gets the next value
void enter_enclosing_container()
{
if (frames.empty())
{
in_object = false;
stored_array = nullptr;
}
else
{
const frame& f = frames.back();
in_object = f.object != nullptr;
stored_array = f.array;
}
}
template<typename Value>
void handle_value(Value&& v)
{
BasicJsonType* const stored = store(std::forward<Value>(v));
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_from_lexer(*stored, m_lexer_ref);
#else
static_cast<void>(stored);
#endif
}
/// the parsed JSON value
BasicJsonType& root;
/// the open arrays and objects
frame_stack frames {};
/// the elements of the open arrays that are not stored in them yet
std::vector<BasicJsonType> elements {};
/// what objects in a vector need besides the frames (std::map based objects
/// need nothing, so they do not carry it)
struct ordered_data_t
{
/// the members of the open objects in a vector that are not stored in them yet
std::vector<std::pair<string_t, BasicJsonType>> members {};
/// the member of the innermost such object that key() added or found last
std::size_t member_index = 0;
/// the key indexes of the open objects that have one, innermost last
std::vector<std::unique_ptr<object_key_index>> key_indexes {};
};
struct no_ordered_data_t {};
typename std::conditional<indexed_keys, ordered_data_t, no_ordered_data_t>::type ordered_data {};
/// helper to hold the reference for the next object element
BasicJsonType* object_element = nullptr;
/// whether the innermost container is an object (frames.back().object)
bool in_object = false;
/// the array that stores the elements of the innermost array as they
/// arrive (frames.back().array), or nullptr
array_t* stored_array = nullptr;
/// whether a syntax error occurred
bool errored = false;
/// whether to throw exceptions in case of errors
const bool allow_exceptions = true;
/// the lexer reference to obtain the current position
lexer_t* m_lexer_ref = nullptr;
};
/*!
@brief SAX implementation to create a JSON value from the events of a binary reader
The DOM parser of the text formats (json_sax_dom_parser) collects the elements
of arrays and the members of objects on a stack and builds them in one
exact-size allocation, which pays off because a text parser does not know the
number of elements up front. Binary formats announce the size of their
containers (CBOR, MessagePack, BSON and BON8 do, and so do the optimized
containers of UBJSON and BJData), so they store each value directly in its
container, and the bookkeeping of the collecting parser would be pure
overhead for them. This is the plain parser: the structure or hierarchy of the
JSON value is managed by the stack `ref_stack` which contains a pointer to the
respective array or object for each recursion depth.
After successful parsing, the value that is passed by reference to the
constructor contains the parsed value.
@tparam BasicJsonType the JSON type
*/
template<typename BasicJsonType, typename InputAdapterType>
class json_sax_dom_binary_parser
{
public:
using number_integer_t = typename BasicJsonType::number_integer_t;
using number_unsigned_t = typename BasicJsonType::number_unsigned_t;
using number_float_t = typename BasicJsonType::number_float_t;
using string_t = typename BasicJsonType::string_t;
using binary_t = typename BasicJsonType::binary_t;
using lexer_t = lexer<BasicJsonType, InputAdapterType>;
/*!
@param[in,out] r reference to a JSON value that is manipulated while
parsing
@param[in] allow_exceptions_ whether parse errors yield exceptions
*/
explicit json_sax_dom_binary_parser(BasicJsonType& r, const bool allow_exceptions_ = true, lexer_t* lexer_ = nullptr)
: root(r), allow_exceptions(allow_exceptions_), m_lexer_ref(lexer_)
{}
// make class move-only
json_sax_dom_binary_parser(const json_sax_dom_binary_parser&) = delete;
json_sax_dom_binary_parser(json_sax_dom_binary_parser&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor)
json_sax_dom_binary_parser& operator=(const json_sax_dom_binary_parser&) = delete;
json_sax_dom_binary_parser& operator=(json_sax_dom_binary_parser&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor)
~json_sax_dom_binary_parser() = default;
bool null()
{
handle_value(nullptr);
return true;
}
bool boolean(bool val)
{
handle_value(val);
return true;
}
bool number_integer(number_integer_t val)
{
handle_value(val);
return true;
}
bool number_unsigned(number_unsigned_t val)
{
handle_value(val);
return true;
}
bool number_float(number_float_t val, const string_t& /*unused*/)
{
handle_value(val);
return true;
}
bool string(string_t& val)
{
// json_sax documents that the passed value may be moved from,
// so hand the buffer over instead of copying it
handle_value(std::move(val));
return true;
}
bool binary(binary_t& val)
{
// json_sax documents that the passed value may be moved from,
// so hand the buffer over instead of copying it
handle_value(std::move(val));
return true;
}
bool start_object(std::size_t len)
{
ref_stack.push_back(handle_value(BasicJsonType::value_t::object));
#if JSON_DIAGNOSTIC_POSITIONS
// Ensure this is after the call to handle_value to ensure correct start position.
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
return check_container_size(*this, len, "object", ref_stack.back());
}
bool key(string_t& val)
{
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(ref_stack.back()->is_object());
// add null at the given key and store the reference for later; the
// key is moved, as the lexer starts the next token with reset()
object_element = &(ref_stack.back()->m_data.m_value.object->operator[](std::move(val)));
return true;
}
bool end_object()
{
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(ref_stack.back()->is_object());
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
ref_stack.pop_back();
return true;
}
bool start_array(std::size_t len)
{
ref_stack.push_back(handle_value(BasicJsonType::value_t::array));
#if JSON_DIAGNOSTIC_POSITIONS
// Ensure this is after the call to handle_value to ensure correct start position.
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
if (JSON_HEDLEY_UNLIKELY(!check_container_size(*this, len, "array", ref_stack.back())))
{
return false;
}
if (len != detail::unknown_size())
{
reserve_array(*ref_stack.back()->m_data.m_value.array, len, priority_tag<1> {});
}
return true;
}
bool end_array()
{
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(ref_stack.back()->is_array());
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
ref_stack.pop_back();
return true;
}
template<class Exception>
bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/,
const Exception& ex)
{
errored = true;
static_cast<void>(ex);
if (allow_exceptions)
{
JSON_THROW(ex);
}
return false;
}
constexpr bool is_errored() const
{
return errored;
}
private:
/*!
@invariant If the ref stack is empty, then the passed value will be the new
root.
@invariant If the ref stack contains a value, then it is an array or an
object to which we can add elements
*/
template<typename Value>
JSON_HEDLEY_RETURNS_NON_NULL
BasicJsonType* handle_value(Value&& v)
{
if (ref_stack.empty())
{
root = BasicJsonType(std::forward<Value>(v));
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_from_lexer(root, m_lexer_ref);
#endif
return &root;
}
JSON_ASSERT(ref_stack.back()->is_array() || ref_stack.back()->is_object());
if (ref_stack.back()->is_array())
{
ref_stack.back()->m_data.m_value.array->emplace_back(std::forward<Value>(v));
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_from_lexer(ref_stack.back()->m_data.m_value.array->back(), m_lexer_ref);
#endif
return &(ref_stack.back()->m_data.m_value.array->back());
}
JSON_ASSERT(ref_stack.back()->is_object());
JSON_ASSERT(object_element);
*object_element = BasicJsonType(std::forward<Value>(v));
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_from_lexer(*object_element, m_lexer_ref);
#endif
return object_element;
}
/// the parsed JSON value
BasicJsonType& root;
/// stack to model hierarchy of values
std::vector<BasicJsonType*> ref_stack {};
/// helper to hold the reference for the next object element
BasicJsonType* object_element = nullptr;
/// whether a syntax error occurred
bool errored = false;
/// whether to throw exceptions in case of errors
const bool allow_exceptions = true;
/// the lexer reference to obtain the current position
lexer_t* m_lexer_ref = nullptr;
};
template<typename BasicJsonType, typename InputAdapterType>
class json_sax_dom_callback_parser
{
public:
using number_integer_t = typename BasicJsonType::number_integer_t;
using number_unsigned_t = typename BasicJsonType::number_unsigned_t;
using number_float_t = typename BasicJsonType::number_float_t;
using string_t = typename BasicJsonType::string_t;
using binary_t = typename BasicJsonType::binary_t;
using parser_callback_t = typename BasicJsonType::parser_callback_t;
using parse_event_t = typename BasicJsonType::parse_event_t;
using lexer_t = lexer<BasicJsonType, InputAdapterType>;
json_sax_dom_callback_parser(BasicJsonType& r,
parser_callback_t cb,
const bool allow_exceptions_ = true,
lexer_t* lexer_ = nullptr)
: root(r), callback(std::move(cb)), allow_exceptions(allow_exceptions_), m_lexer_ref(lexer_)
{
keep_stack.push_back(true);
}
// make class move-only
json_sax_dom_callback_parser(const json_sax_dom_callback_parser&) = delete;
json_sax_dom_callback_parser(json_sax_dom_callback_parser&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor)
json_sax_dom_callback_parser& operator=(const json_sax_dom_callback_parser&) = delete;
json_sax_dom_callback_parser& operator=(json_sax_dom_callback_parser&&) = default; // NOLINT(hicpp-noexcept-move,performance-noexcept-move-constructor)
~json_sax_dom_callback_parser() = default;
bool null()
{
handle_value(nullptr);
return true;
}
bool boolean(bool val)
{
handle_value(val);
return true;
}
bool number_integer(number_integer_t val)
{
handle_value(val);
return true;
}
bool number_unsigned(number_unsigned_t val)
{
handle_value(val);
return true;
}
bool number_float(number_float_t val, const string_t& /*unused*/)
{
handle_value(val);
return true;
}
bool string(string_t& val)
{
// json_sax documents that the passed value may be moved from,
// so hand the buffer over instead of copying it
handle_value(std::move(val));
return true;
}
bool binary(binary_t& val)
{
// json_sax documents that the passed value may be moved from,
// so hand the buffer over instead of copying it
handle_value(std::move(val));
return true;
}
bool start_object(std::size_t len)
{
// check callback for object start; not called inside a discarded container
const bool keep = keep_stack.back() && callback(static_cast<int>(ref_stack.size()), parse_event_t::object_start, discarded);
keep_stack.push_back(keep);
// the key this object will be stored under, read before handle_value()
// may consume it; kept in lockstep with ref_stack so end_object() can
// find the object in its parent again
container_key_stack.push_back(current_key());
auto val = handle_value(BasicJsonType::value_t::object, true);
ref_stack.push_back(val.second);
if (ref_stack.back())
{
#if JSON_DIAGNOSTIC_POSITIONS
// Ensure this is after the call to handle_value to ensure correct start position.
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
return check_container_size(*this, len, "object", ref_stack.back());
}
return true;
}
bool key(string_t& val)
{
if (!keep_stack.back() || !ref_stack.back())
{
// the object is not stored: the value of this key is dropped in
// handle_value() without touching the key stacks
if (keep_stack.back())
{
static_cast<void>(callback(static_cast<int>(ref_stack.size()), parse_event_t::key, key_value(val)));
}
return true;
}
// check callback for the key
const bool keep = callback(static_cast<int>(ref_stack.size()), parse_event_t::key, key_value(val));
key_keep_stack.push_back(keep);
// remember the key so a rejected value can be erased without searching
// the object for it (kept in lockstep with key_keep_stack)
key_stack.push_back(val);
// add discarded value at the given key and store the reference for later
if (keep && ref_stack.back())
{
auto& obj = *ref_stack.back()->m_data.m_value.object;
// a single lookup: insert the placeholder, or find the existing member
const auto inserted = obj.emplace(std::move(val), discarded);
const auto it = inserted.first;
if (!inserted.second)
{
// this is a duplicate key (legal in JSON); remember its
// current value so it can be restored later if the new
// value is rejected by the callback, instead of being
// erased together with the discarded placeholder
duplicate_key_stash.emplace_back(&(it->second), it->second);
it->second = discarded;
}
object_element = &(it->second);
}
return true;
}
bool end_object()
{
if (ref_stack.back())
{
if (!callback(static_cast<int>(ref_stack.size()) - 1, parse_event_t::object_end, *ref_stack.back()))
{
// discard object, unless this slot holds a duplicate key's
// previous value pending restoration, in which case that
// value is restored instead of being discarded
if (!resolve_duplicate_key_stash(ref_stack.back(), true))
{
*ref_stack.back() = discarded;
#if JSON_DIAGNOSTIC_POSITIONS
// Set start/end positions for discarded object.
diagnostic_positions::set_from_lexer(*ref_stack.back(), m_lexer_ref);
#endif
}
}
else
{
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
// this object is finally, definitively kept; drop any
// pending duplicate-key stash entry for its slot since it
// can no longer be restored
resolve_duplicate_key_stash(ref_stack.back(), false);
}
}
const string_t object_key = pop_container();
if (!ref_stack.empty() && ref_stack.back() && ref_stack.back()->is_structured())
{
// remove discarded value
remove_discarded_value(*ref_stack.back(), object_key);
}
return true;
}
bool start_array(std::size_t len)
{
const bool keep = keep_stack.back() && callback(static_cast<int>(ref_stack.size()), parse_event_t::array_start, discarded);
keep_stack.push_back(keep);
// see start_object()
container_key_stack.push_back(current_key());
auto val = handle_value(BasicJsonType::value_t::array, true);
ref_stack.push_back(val.second);
if (ref_stack.back())
{
#if JSON_DIAGNOSTIC_POSITIONS
// Ensure this is after the call to handle_value to ensure correct start position.
diagnostic_positions::set_container_start(*ref_stack.back(), m_lexer_ref);
#endif
if (JSON_HEDLEY_UNLIKELY(!check_container_size(*this, len, "array", ref_stack.back())))
{
return false;
}
if (len != detail::unknown_size())
{
reserve_array(*ref_stack.back()->m_data.m_value.array, len, priority_tag<1> {});
}
}
return true;
}
bool end_array()
{
bool keep = true;
const bool stored = ref_stack.back() != nullptr;
if (stored)
{
keep = callback(static_cast<int>(ref_stack.size()) - 1, parse_event_t::array_end, *ref_stack.back());
if (keep)
{
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_container_end(*ref_stack.back(), m_lexer_ref);
#endif
ref_stack.back()->set_parents();
// this array is finally, definitively kept; drop any
// pending duplicate-key stash entry for its slot since it
// can no longer be restored
resolve_duplicate_key_stash(ref_stack.back(), false);
}
else
{
// discard array, unless this slot holds a duplicate key's
// previous value pending restoration, in which case that
// value is restored instead of being discarded
if (!resolve_duplicate_key_stash(ref_stack.back(), true))
{
*ref_stack.back() = discarded;
#if JSON_DIAGNOSTIC_POSITIONS
// Set start/end positions for discarded array.
diagnostic_positions::set_from_lexer(*ref_stack.back(), m_lexer_ref);
#endif
}
}
}
const string_t object_key = pop_container();
// remove discarded value
if (!ref_stack.empty() && ref_stack.back())
{
if (!keep && ref_stack.back()->is_array())
{
ref_stack.back()->m_data.m_value.array->pop_back();
}
else if ((!keep || !stored) && ref_stack.back()->is_object())
{
// the array is either still stored under its key or was never
// stored, leaving the placeholder key() wrote; both show up as
// a discarded member of the parent object
remove_discarded_value(*ref_stack.back(), object_key);
}
}
return true;
}
template<class Exception>
bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/,
const Exception& ex)
{
errored = true;
static_cast<void>(ex);
if (allow_exceptions)
{
JSON_THROW(ex);
}
return false;
}
constexpr bool is_errored() const
{
return errored;
}
private:
/*!
@brief the value of a key event for the callback
A string value of its own for every key would allocate it every time, so
one is kept and set to the key. A callback that changed or moved it gets a
new one, as it would have got a new value for every key anyway.
*/
BasicJsonType& key_value(const string_t& key)
{
#if JSON_DIAGNOSTIC_POSITIONS
// a callback may have left positions on the value
key_scratch = BasicJsonType(key);
#else
if (JSON_HEDLEY_LIKELY(key_scratch.is_string()))
{
*key_scratch.m_data.m_value.string = key;
}
else
{
key_scratch = BasicJsonType(key);
}
#endif
return key_scratch;
}
/// if there is a pending duplicate-key stash entry for this exact slot,
/// remove it from the stash; if restore_value is true, the stashed
/// previous value is moved back into the slot first (use this when the
/// new value at that slot was rejected); otherwise the stash entry is
/// simply dropped (use this when the new value was accepted, so it
/// correctly supersedes the old one and no restore should ever happen
/// for this slot again)
/// @return whether a matching stash entry was found (and processed)
bool resolve_duplicate_key_stash(BasicJsonType* slot, bool restore_value)
{
const auto it = std::find_if(duplicate_key_stash.begin(), duplicate_key_stash.end(),
[slot](const std::pair<BasicJsonType*, BasicJsonType>& entry)
{
return entry.first == slot;
});
if (it == duplicate_key_stash.end())
{
return false;
}
if (restore_value)
{
*slot = std::move(it->second);
}
duplicate_key_stash.erase(it);
return true;
}
/*!
@brief the key the value now being handled will be stored under
Empty unless the enclosing container is an object, in which case it is the
key of the pending key() event. Read before handle_value() consumes that
key, so it is also correct when the value never reaches its parent.
*/
string_t current_key() const
{
if (!ref_stack.empty() && ref_stack.back() && ref_stack.back()->is_object()
&& !key_stack.empty())
{
return key_stack.back();
}
return string_t{};
}
/*!
@brief leave the object or array that end_object()/end_array() just closed
@return the key the container is stored under in its parent (see
container_key_stack)
*/
string_t pop_container()
{
JSON_ASSERT(!ref_stack.empty());
JSON_ASSERT(!keep_stack.empty());
JSON_ASSERT(!container_key_stack.empty());
ref_stack.pop_back();
keep_stack.pop_back();
string_t object_key = std::move(container_key_stack.back());
container_key_stack.pop_back();
return object_key;
}
/*!
@brief remove the discarded value the callback rejected from its parent,
unless it is a duplicate key's slot with a stashed previous value, in
which case that previous value is restored instead
A rejected value can only ever be the one most recently added to @a parent:
the last element of an array, or the placeholder key() stored under @a key
in an object. Looking there directly makes this O(1) resp. O(log n), where
searching @a parent for it made a filtering parse quadratic in the number of
members of a single container.
Finding no discarded value there means none was stored in the first place -
the callback rejected the value before it reached its parent - so there is
nothing to remove.
@param[in,out] parent the container to remove the rejected value from
@param[in] key the key the value was stored under; unused for arrays
*/
void remove_discarded_value(BasicJsonType& parent, const string_t& key)
{
if (parent.is_array())
{
auto& array = *parent.m_data.m_value.array;
if (!array.empty() && array.back().is_discarded())
{
array.pop_back();
}
}
else if (parent.is_object())
{
auto& object = *parent.m_data.m_value.object;
const auto it = object.find(key);
if (it != object.end() && it->second.is_discarded())
{
// a duplicate key's slot has a stashed previous value that
// must be restored instead of being erased
if (!resolve_duplicate_key_stash(&it->second, true))
{
object.erase(it);
}
}
}
}
/*!
@param[in] v value to add to the JSON value we build during parsing
@param[in] skip_callback whether we should skip calling the callback
function; this is required after start_array() and
start_object() SAX events, because otherwise we would call the
callback function with an empty array or object, respectively.
@invariant If the ref stack is empty, then the passed value will be the new
root.
@invariant If the ref stack contains a value, then it is an array or an
object to which we can add elements
@return pair of boolean (whether value should be kept) and pointer (to the
passed value in the ref_stack hierarchy; nullptr if not kept)
*/
template<typename Value>
std::pair<bool, BasicJsonType*> handle_value(Value&& v, const bool skip_callback = false)
{
JSON_ASSERT(!keep_stack.empty());
// do not handle this value if we know it would be added to a discarded
// container
if (!keep_stack.back())
{
return {false, nullptr};
}
// create value
auto value = BasicJsonType(std::forward<Value>(v));
#if JSON_DIAGNOSTIC_POSITIONS
diagnostic_positions::set_from_lexer(value, m_lexer_ref);
#endif
// check callback
const bool keep = skip_callback || callback(static_cast<int>(ref_stack.size()), parse_event_t::value, value);
// do not handle this value if we just learnt it shall be discarded
if (!keep)
{
// if the value was to become an object member, key() already
// stored a placeholder for it that has to be removed again
if (!ref_stack.empty() && ref_stack.back() && ref_stack.back()->is_object())
{
JSON_ASSERT(!key_keep_stack.empty());
JSON_ASSERT(!key_stack.empty());
const bool placeholder_stored = key_keep_stack.back();
key_keep_stack.pop_back();
const string_t key = std::move(key_stack.back());
key_stack.pop_back();
if (placeholder_stored)
{
remove_discarded_value(*ref_stack.back(), key);
}
}
return {false, nullptr};
}
if (ref_stack.empty())
{
root = std::move(value);
return {true, & root};
}
// skip this value if we already decided to skip the parent
// (https://github.com/nlohmann/json/issues/971#issuecomment-413678360)
if (!ref_stack.back())
{
return {false, nullptr};
}
// we now only expect arrays and objects
JSON_ASSERT(ref_stack.back()->is_array() || ref_stack.back()->is_object());
// array
if (ref_stack.back()->is_array())
{
ref_stack.back()->m_data.m_value.array->emplace_back(std::move(value));
return {true, & (ref_stack.back()->m_data.m_value.array->back())};
}
// object
JSON_ASSERT(ref_stack.back()->is_object());
// check if we should store an element for the current key
JSON_ASSERT(!key_keep_stack.empty());
JSON_ASSERT(!key_stack.empty());
const bool store_element = key_keep_stack.back();
key_keep_stack.pop_back();
key_stack.pop_back();
if (!store_element)
{
return {false, nullptr};
}
JSON_ASSERT(object_element);
*object_element = std::move(value);
if (!skip_callback)
{
// this scalar value finally, definitively replaces whatever was
// at this slot; drop any pending duplicate-key stash entry for
// it since it can no longer be restored (a container value at
// this slot is resolved later, in end_object()/end_array(),
// since skip_callback is true for the placeholder handling that
// happens here for those)
resolve_duplicate_key_stash(object_element, false);
}
return {true, object_element};
}
/// the parsed JSON value
BasicJsonType& root;
/// stack to model hierarchy of values
std::vector<BasicJsonType*> ref_stack {};
/// stack to manage which values to keep
std::vector<bool> keep_stack {}; // NOLINT(readability-redundant-member-init)
/// stack to manage which object keys to keep
std::vector<bool> key_keep_stack {}; // NOLINT(readability-redundant-member-init)
/// the keys key() stored a placeholder for, in lockstep with key_keep_stack
std::vector<string_t> key_stack {}; // NOLINT(readability-redundant-member-init)
/// for each open container, the key it is stored under in its parent
/// object, in lockstep with ref_stack; unused where the parent is not an
/// object
std::vector<string_t> container_key_stack {}; // NOLINT(readability-redundant-member-init)
/// helper to hold the reference for the next object element
BasicJsonType* object_element = nullptr;
/// stash of (slot pointer, previous value) for object members that
/// already existed when key() was called again for the same key
/// (duplicate keys); used to restore the previous value if the new
/// value is later rejected by the callback, instead of erasing the
/// member entirely
std::vector<std::pair<BasicJsonType*, BasicJsonType>> duplicate_key_stash {};
/// whether a syntax error occurred
bool errored = false;
/// callback function
const parser_callback_t callback = nullptr;
/// whether to throw exceptions in case of errors
const bool allow_exceptions = true;
/// a discarded value for the callback
BasicJsonType discarded = BasicJsonType::value_t::discarded;
/// the value of the key event for the callback (see key_value())
BasicJsonType key_scratch {};
/// the lexer reference to obtain the current position
lexer_t* m_lexer_ref = nullptr;
};
template<typename BasicJsonType>
class json_sax_acceptor
{
public:
using number_integer_t = typename BasicJsonType::number_integer_t;
using number_unsigned_t = typename BasicJsonType::number_unsigned_t;
using number_float_t = typename BasicJsonType::number_float_t;
using string_t = typename BasicJsonType::string_t;
using binary_t = typename BasicJsonType::binary_t;
bool null()
{
return true;
}
bool boolean(bool /*unused*/)
{
return true;
}
bool number_integer(number_integer_t /*unused*/)
{
return true;
}
bool number_unsigned(number_unsigned_t /*unused*/)
{
return true;
}
bool number_float(number_float_t /*unused*/, const string_t& /*unused*/)
{
return true;
}
bool string(string_t& /*unused*/)
{
return true;
}
bool binary(binary_t& /*unused*/)
{
return true;
}
bool start_object(std::size_t /*unused*/ = detail::unknown_size())
{
return true;
}
bool key(string_t& /*unused*/)
{
return true;
}
bool end_object()
{
return true;
}
bool start_array(std::size_t /*unused*/ = detail::unknown_size())
{
return true;
}
bool end_array()
{
return true;
}
bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/, const detail::exception& /*unused*/)
{
return false;
}
};
} // namespace detail
NLOHMANN_JSON_NAMESPACE_END
Reference in new issue View Git Blame Copy Permalink