diff options
Diffstat (limited to 'include/lib/modernjson/detail/input/parser.hpp')
| -rw-r--r-- | include/lib/modernjson/detail/input/parser.hpp | 488 |
1 files changed, 488 insertions, 0 deletions
diff --git a/include/lib/modernjson/detail/input/parser.hpp b/include/lib/modernjson/detail/input/parser.hpp new file mode 100644 index 0000000..a622077 --- /dev/null +++ b/include/lib/modernjson/detail/input/parser.hpp @@ -0,0 +1,488 @@ +#pragma once + +#include <cassert> // assert +#include <cmath> // isfinite +#include <cstdint> // uint8_t +#include <functional> // function +#include <string> // string +#include <utility> // move + +#include <lib/modernjson/detail/exceptions.hpp> +#include <lib/modernjson/detail/macro_scope.hpp> +#include <lib/modernjson/detail/meta/is_sax.hpp> +#include <lib/modernjson/detail/input/input_adapters.hpp> +#include <lib/modernjson/detail/input/json_sax.hpp> +#include <lib/modernjson/detail/input/lexer.hpp> +#include <lib/modernjson/detail/value_t.hpp> + +namespace nlohmann +{ +namespace detail +{ +//////////// +// parser // +//////////// + +/*! +@brief syntax analysis + +This class implements a recursive decent parser. +*/ +template<typename BasicJsonType> +class parser +{ + 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 lexer_t = lexer<BasicJsonType>; + using token_type = typename lexer_t::token_type; + + public: + enum class parse_event_t : uint8_t + { + /// the parser read `{` and started to process a JSON object + object_start, + /// the parser read `}` and finished processing a JSON object + object_end, + /// the parser read `[` and started to process a JSON array + array_start, + /// the parser read `]` and finished processing a JSON array + array_end, + /// the parser read a key of a value in an object + key, + /// the parser finished reading a JSON value + value + }; + + using parser_callback_t = + std::function<bool(int depth, parse_event_t event, BasicJsonType& parsed)>; + + /// a parser reading from an input adapter + explicit parser(detail::input_adapter_t&& adapter, + const parser_callback_t cb = nullptr, + const bool allow_exceptions_ = true) + : callback(cb), m_lexer(std::move(adapter)), allow_exceptions(allow_exceptions_) + { + // read first token + get_token(); + } + + /*! + @brief public parser interface + + @param[in] strict whether to expect the last token to be EOF + @param[in,out] result parsed JSON value + + @throw parse_error.101 in case of an unexpected token + @throw parse_error.102 if to_unicode fails or surrogate error + @throw parse_error.103 if to_unicode fails + */ + void parse(const bool strict, BasicJsonType& result) + { + if (callback) + { + json_sax_dom_callback_parser<BasicJsonType> sdp(result, callback, allow_exceptions); + sax_parse_internal(&sdp); + result.assert_invariant(); + + // in strict mode, input must be completely read + if (strict and (get_token() != token_type::end_of_input)) + { + sdp.parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::end_of_input))); + } + + // in case of an error, return discarded value + if (sdp.is_errored()) + { + result = value_t::discarded; + return; + } + + // set top-level value to null if it was discarded by the callback + // function + if (result.is_discarded()) + { + result = nullptr; + } + } + else + { + json_sax_dom_parser<BasicJsonType> sdp(result, allow_exceptions); + sax_parse_internal(&sdp); + result.assert_invariant(); + + // in strict mode, input must be completely read + if (strict and (get_token() != token_type::end_of_input)) + { + sdp.parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::end_of_input))); + } + + // in case of an error, return discarded value + if (sdp.is_errored()) + { + result = value_t::discarded; + return; + } + } + } + + /*! + @brief public accept interface + + @param[in] strict whether to expect the last token to be EOF + @return whether the input is a proper JSON text + */ + bool accept(const bool strict = true) + { + json_sax_acceptor<BasicJsonType> sax_acceptor; + return sax_parse(&sax_acceptor, strict); + } + + template <typename SAX> + bool sax_parse(SAX* sax, const bool strict = true) + { + (void)detail::is_sax_static_asserts<SAX, BasicJsonType> {}; + const bool result = sax_parse_internal(sax); + + // strict mode: next byte must be EOF + if (result and strict and (get_token() != token_type::end_of_input)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::end_of_input))); + } + + return result; + } + + private: + template <typename SAX> + bool sax_parse_internal(SAX* sax) + { + // stack to remember the hieararchy of structured values we are parsing + // true = array; false = object + std::vector<bool> states; + // value to avoid a goto (see comment where set to true) + bool skip_to_state_evaluation = false; + + while (true) + { + if (not skip_to_state_evaluation) + { + // invariant: get_token() was called before each iteration + switch (last_token) + { + case token_type::begin_object: + { + if (JSON_UNLIKELY(not sax->start_object(std::size_t(-1)))) + { + return false; + } + + // closing } -> we are done + if (get_token() == token_type::end_object) + { + if (JSON_UNLIKELY(not sax->end_object())) + { + return false; + } + break; + } + + // parse key + if (JSON_UNLIKELY(last_token != token_type::value_string)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::value_string))); + } + else + { + if (JSON_UNLIKELY(not sax->key(m_lexer.get_string()))) + { + return false; + } + } + + // parse separator (:) + if (JSON_UNLIKELY(get_token() != token_type::name_separator)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::name_separator))); + } + + // remember we are now inside an object + states.push_back(false); + + // parse values + get_token(); + continue; + } + + case token_type::begin_array: + { + if (JSON_UNLIKELY(not sax->start_array(std::size_t(-1)))) + { + return false; + } + + // closing ] -> we are done + if (get_token() == token_type::end_array) + { + if (JSON_UNLIKELY(not sax->end_array())) + { + return false; + } + break; + } + + // remember we are now inside an array + states.push_back(true); + + // parse values (no need to call get_token) + continue; + } + + case token_type::value_float: + { + const auto res = m_lexer.get_number_float(); + + if (JSON_UNLIKELY(not std::isfinite(res))) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + out_of_range::create(406, "number overflow parsing '" + m_lexer.get_token_string() + "'")); + } + else + { + if (JSON_UNLIKELY(not sax->number_float(res, m_lexer.get_string()))) + { + return false; + } + break; + } + } + + case token_type::literal_false: + { + if (JSON_UNLIKELY(not sax->boolean(false))) + { + return false; + } + break; + } + + case token_type::literal_null: + { + if (JSON_UNLIKELY(not sax->null())) + { + return false; + } + break; + } + + case token_type::literal_true: + { + if (JSON_UNLIKELY(not sax->boolean(true))) + { + return false; + } + break; + } + + case token_type::value_integer: + { + if (JSON_UNLIKELY(not sax->number_integer(m_lexer.get_number_integer()))) + { + return false; + } + break; + } + + case token_type::value_string: + { + if (JSON_UNLIKELY(not sax->string(m_lexer.get_string()))) + { + return false; + } + break; + } + + case token_type::value_unsigned: + { + if (JSON_UNLIKELY(not sax->number_unsigned(m_lexer.get_number_unsigned()))) + { + return false; + } + break; + } + + case token_type::parse_error: + { + // using "uninitialized" to avoid "expected" message + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::uninitialized))); + } + + default: // the last token was unexpected + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::literal_or_value))); + } + } + } + else + { + skip_to_state_evaluation = false; + } + + // we reached this line after we successfully parsed a value + if (states.empty()) + { + // empty stack: we reached the end of the hieararchy: done + return true; + } + else + { + if (states.back()) // array + { + // comma -> next value + if (get_token() == token_type::value_separator) + { + // parse a new value + get_token(); + continue; + } + + // closing ] + if (JSON_LIKELY(last_token == token_type::end_array)) + { + if (JSON_UNLIKELY(not sax->end_array())) + { + return false; + } + + // We are done with this array. Before we can parse a + // new value, we need to evaluate the new state first. + // By setting skip_to_state_evaluation to false, we + // are effectively jumping to the beginning of this if. + assert(not states.empty()); + states.pop_back(); + skip_to_state_evaluation = true; + continue; + } + else + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::end_array))); + } + } + else // object + { + // comma -> next value + if (get_token() == token_type::value_separator) + { + // parse key + if (JSON_UNLIKELY(get_token() != token_type::value_string)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::value_string))); + } + else + { + if (JSON_UNLIKELY(not sax->key(m_lexer.get_string()))) + { + return false; + } + } + + // parse separator (:) + if (JSON_UNLIKELY(get_token() != token_type::name_separator)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::name_separator))); + } + + // parse values + get_token(); + continue; + } + + // closing } + if (JSON_LIKELY(last_token == token_type::end_object)) + { + if (JSON_UNLIKELY(not sax->end_object())) + { + return false; + } + + // We are done with this object. Before we can parse a + // new value, we need to evaluate the new state first. + // By setting skip_to_state_evaluation to false, we + // are effectively jumping to the beginning of this if. + assert(not states.empty()); + states.pop_back(); + skip_to_state_evaluation = true; + continue; + } + else + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), exception_message(token_type::end_object))); + } + } + } + } + } + + /// get next token from lexer + token_type get_token() + { + return (last_token = m_lexer.scan()); + } + + std::string exception_message(const token_type expected) + { + std::string error_msg = "syntax error - "; + if (last_token == token_type::parse_error) + { + error_msg += std::string(m_lexer.get_error_message()) + "; last read: '" + + m_lexer.get_token_string() + "'"; + } + else + { + error_msg += "unexpected " + std::string(lexer_t::token_type_name(last_token)); + } + + if (expected != token_type::uninitialized) + { + error_msg += "; expected " + std::string(lexer_t::token_type_name(expected)); + } + + return error_msg; + } + + private: + /// callback function + const parser_callback_t callback = nullptr; + /// the type of the last read token + token_type last_token = token_type::uninitialized; + /// the lexer + lexer_t m_lexer; + /// whether to throw exceptions in case of errors + const bool allow_exceptions = true; +}; +} +} |