# HG changeset patch # User David Demelier # Date 1475663922 -7200 # Node ID 69cf95733eea89a95c6247b1c3ec7f5c0c1da540 # Parent ea209f52628150c3ea112c71ceaeecc0be5a08c2 json: upgrade to 2.0.5 diff -r ea209f526281 -r 69cf95733eea VERSION.json.txt --- a/VERSION.json.txt Sun Sep 04 10:45:52 2016 +0200 +++ b/VERSION.json.txt Wed Oct 05 12:38:42 2016 +0200 @@ -1,1 +1,1 @@ -2.0.0 +2.0.5 diff -r ea209f526281 -r 69cf95733eea json/json.hpp --- a/json/json.hpp Sun Sep 04 10:45:52 2016 +0200 +++ b/json/json.hpp Wed Oct 05 12:38:42 2016 +0200 @@ -1,7 +1,7 @@ /* __ _____ _____ _____ __| | __| | | | JSON for Modern C++ -| | |__ | | | | | | version 2.0.1 +| | |__ | | | | | | version 2.0.5 |_____|_____|_____|_|___| https://github.com/nlohmann/json Licensed under the MIT License . @@ -32,20 +32,22 @@ #include #include #include -#include #include #include #include -#include +#include #include +#include #include #include #include #include #include #include +#include #include #include +#include #include #include #include @@ -53,12 +55,34 @@ #include #include +// exclude unsupported compilers +#if defined(__clang__) + #define CLANG_VERSION (__clang_major__ * 10000 + __clang_minor__ * 100 + __clang_patchlevel__) + #if CLANG_VERSION < 30400 + #error "unsupported Clang version - see https://github.com/nlohmann/json#supported-compilers" + #endif +#elif defined(__GNUC__) + #define GCC_VERSION (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) + #if GCC_VERSION < 40900 + #error "unsupported GCC version - see https://github.com/nlohmann/json#supported-compilers" + #endif +#endif + // disable float-equal warnings on GCC/clang #if defined(__clang__) || defined(__GNUC__) || defined(__GNUG__) #pragma GCC diagnostic push #pragma GCC diagnostic ignored "-Wfloat-equal" #endif +// allow for portable deprecation warnings +#if defined(__clang__) || defined(__GNUC__) || defined(__GNUG__) + #define JSON_DEPRECATED __attribute__((deprecated)) +#elif defined(_MSC_VER) + #define JSON_DEPRECATED __declspec(deprecated) +#else + #define JSON_DEPRECATED +#endif + /*! @brief namespace for Niels Lohmann @see https://github.com/nlohmann @@ -76,7 +100,13 @@ { /*! @brief Helper to determine whether there's a key_type for T. + +Thus helper is used to tell associative containers apart from other containers +such as sequence containers. For instance, `std::map` passes the test as it +contains a `mapped_type`, whereas `std::vector` fails the test. + @sa http://stackoverflow.com/a/7728728/266378 +@since version 1.0.0 */ template struct has_mapped_type @@ -90,11 +120,18 @@ /*! @brief helper class to create locales with decimal point + +This struct is used a default locale during the JSON serialization. JSON +requires the decimal point to be `.`, so this function overloads the +`do_decimal_point()` function to return `.`. This function is called by +float-to-string conversions to retrieve the decimal separator between integer +and fractional parts. + @sa https://github.com/nlohmann/json/issues/51#issuecomment-86869315 +@since version 2.0.0 */ -class DecimalSeparator : public std::numpunct +struct DecimalSeparator : std::numpunct { - protected: char do_decimal_point() const { return '.'; @@ -163,6 +200,13 @@ JSON values can be used like STL containers and provide reverse iterator access. +@invariant The member variables @a m_value and @a m_type have the following +relationship: +- If `m_type == value_t::object`, then `m_value.object != nullptr`. +- If `m_type == value_t::array`, then `m_value.array != nullptr`. +- If `m_type == value_t::string`, then `m_value.string != nullptr`. +The invariants are checked by member function assert_invariant(). + @internal @note ObjectType trick from http://stackoverflow.com/a/9860911 @endinternal @@ -188,13 +232,8 @@ { private: /// workaround type for MSVC - using basic_json_t = basic_json; public: @@ -207,6 +246,8 @@ ///////////////////// /// @name container types + /// The canonic container types to use @ref basic_json like any other STL + /// container. /// @{ /// the type of elements in a basic_json container @@ -256,6 +297,8 @@ /////////////////////////// /// @name JSON value data types + /// The data types to store a JSON value. These types are derived from + /// the template arguments passed to class @ref basic_json. /// @{ /*! @@ -358,7 +401,7 @@ @tparam ArrayType container type to store arrays (e.g., `std::vector` or `std::list`) - @tparam AllocatorType allocator to use for arrays (e.g., `std::allocator`) + @tparam AllocatorType allocator to use for arrays (e.g., `std::allocator`) #### Default type @@ -595,15 +638,14 @@ > that implementations will agree exactly on their numeric values. As this range is a subrange (when considered in conjunction with the - number_integer_t type) of the exactly supported range [0, UINT64_MAX], this - class's integer type is interoperable. + number_integer_t type) of the exactly supported range [0, UINT64_MAX], + this class's integer type is interoperable. #### Storage Integer number values are stored directly inside a @ref basic_json type. @sa @ref number_float_t -- type for number values (floating-point) - @sa @ref number_integer_t -- type for number values (integer) @since version 2.0.0 @@ -691,7 +733,19 @@ This enumeration collects the different JSON types. It is internally used to distinguish the stored values, and the functions @ref is_null(), @ref is_object(), @ref is_array(), @ref is_string(), @ref is_boolean(), @ref - is_number(), and @ref is_discarded() rely on it. + is_number() (with @ref is_number_integer(), @ref is_number_unsigned(), and + @ref is_number_float()), @ref is_discarded(), @ref is_primitive(), and + @ref is_structured() rely on it. + + @note There are three enumeration entries (number_integer, + number_unsigned, and number_float), because the library distinguishes + these three types for numbers: @ref number_unsigned_t is used for unsigned + integers, @ref number_integer_t is used for signed integers, and @ref + number_float_t is used for floating-point numbers or to approximate + integers which do not fit in the limits of their respective type. + + @sa @ref basic_json(const value_t value_type) -- create a JSON value with + the default value for a given type @since version 1.0.0 */ @@ -702,7 +756,7 @@ array, ///< array (ordered collection of values) string, ///< string value boolean, ///< boolean value - number_integer, ///< number value (integer) + number_integer, ///< number value (signed integer) number_unsigned, ///< number value (unsigned integer) number_float, ///< number value (floating-point) discarded ///< discarded by the the parser callback function @@ -722,6 +776,7 @@ }; std::unique_ptr object(alloc.allocate(1), deleter); alloc.construct(object.get(), std::forward(args)...); + assert(object.get() != nullptr); return object.release(); } @@ -732,7 +787,24 @@ /*! @brief a JSON value - The actual storage for a JSON value of the @ref basic_json class. + The actual storage for a JSON value of the @ref basic_json class. This + union combines the different storage types for the JSON value types + defined in @ref value_t. + + JSON type | value_t type | used type + --------- | --------------- | ------------------------ + object | object | pointer to @ref object_t + array | array | pointer to @ref array_t + string | string | pointer to @ref string_t + boolean | boolean | @ref boolean_t + number | number_integer | @ref number_integer_t + number | number_unsigned | @ref number_unsigned_t + number | number_float | @ref number_float_t + null | null | *no value is stored* + + @note Variable-length types (objects, arrays, and strings) are stored as + pointers. The size of the union should not exceed 64 bits if the default + value types are used. @since version 1.0.0 */ @@ -836,6 +908,21 @@ } }; + /*! + @brief checks the class invariants + + This function asserts the class invariants. It needs to be called at the + end of every constructor to make sure that created objects respect the + invariant. Furthermore, it has to be called each time the type of a JSON + value is changed, because the invariant expresses a relationship between + @a m_type and @a m_value. + */ + void assert_invariant() const + { + assert(m_type != value_t::object or m_value.object != nullptr); + assert(m_type != value_t::array or m_value.array != nullptr); + assert(m_type != value_t::string or m_value.string != nullptr); + } public: ////////////////////////// @@ -848,6 +935,8 @@ This enumeration lists the parser events that can trigger calling a callback function of type @ref parser_callback_t during parsing. + @image html callback_events.png "Example when certain parse events are triggered" + @since version 1.0.0 */ enum class parse_event_t : uint8_t @@ -870,12 +959,13 @@ @brief per-element parser callback type With a parser callback function, the result of parsing a JSON text can be - influenced. When passed to @ref parse(std::istream&, parser_callback_t) or - @ref parse(const string_t&, parser_callback_t), it is called on certain - events (passed as @ref parse_event_t via parameter @a event) with a set - recursion depth @a depth and context JSON value @a parsed. The return - value of the callback function is a boolean indicating whether the element - that emitted the callback shall be kept or not. + influenced. When passed to @ref parse(std::istream&, const + parser_callback_t) or @ref parse(const char*, const parser_callback_t), + it is called on certain events (passed as @ref parse_event_t via parameter + @a event) with a set recursion depth @a depth and context JSON value + @a parsed. The return value of the callback function is a boolean + indicating whether the element that emitted the callback shall be kept or + not. We distinguish six scenarios (determined by the event type) in which the callback function can be called. The following table describes the values @@ -890,6 +980,8 @@ parse_event_t::array_end | the parser read `]` and finished processing a JSON array | depth of the parent of the JSON array | the parsed JSON array parse_event_t::value | the parser finished reading a JSON value | depth of the value | the parsed JSON value + @image html callback_events.png "Example when certain parse events are triggered" + Discarding a value (i.e., returning `false`) has different effects depending on the context in which function was called: @@ -911,11 +1003,13 @@ skipped completely or replaced by an empty discarded object. @sa @ref parse(std::istream&, parser_callback_t) or - @ref parse(const string_t&, parser_callback_t) for examples - - @since version 1.0.0 - */ - using parser_callback_t = std::function; + @ref parse(const char*, parser_callback_t) for examples + + @since version 1.0.0 + */ + using parser_callback_t = std::function; ////////////////// @@ -923,6 +1017,8 @@ ////////////////// /// @name constructors and destructors + /// Constructors of class @ref basic_json, copy/move constructor, copy + /// assignment, static functions creating objects, and the destructor. /// @{ /*! @@ -966,40 +1062,15 @@ */ basic_json(const value_t value_type) : m_type(value_type), m_value(value_type) - {} - - /*! - @brief create a null object (implicitly) - - Create a `null` JSON value. This is the implicit version of the `null` - value constructor as it takes no parameters. - - @complexity Constant. - - @exceptionsafety No-throw guarantee: this constructor never throws - exceptions. - - @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) - requirements: - - The complexity is constant. - - As postcondition, it holds: `basic_json().empty() == true`. - - @liveexample{The following code shows the constructor for a `null` JSON - value.,basic_json} - - @sa @ref basic_json(std::nullptr_t) -- create a `null` value - - @since version 1.0.0 - */ - basic_json() = default; - - /*! - @brief create a null object (explicitly) - - Create a `null` JSON value. This is the explicitly version of the `null` - value constructor as it takes a null pointer as parameter. It allows to - create `null` values by explicitly assigning a `nullptr` to a JSON value. + { + assert_invariant(); + } + + /*! + @brief create a null object + + Create a `null` JSON value. It either takes a null pointer as parameter + (explicitly creating `null`) or no parameter (implicitly creating `null`). The passed null pointer itself is not read -- it is only used to choose the right constructor. @@ -1008,17 +1079,16 @@ @exceptionsafety No-throw guarantee: this constructor never throws exceptions. - @liveexample{The following code shows the constructor with null pointer - parameter.,basic_json__nullptr_t} - - @sa @ref basic_json() -- default constructor (implicitly creating a `null` - value) - - @since version 1.0.0 - */ - basic_json(std::nullptr_t) noexcept + @liveexample{The following code shows the constructor with and without a + null pointer parameter.,basic_json__nullptr_t} + + @since version 1.0.0 + */ + basic_json(std::nullptr_t = nullptr) noexcept : basic_json(value_t::null) - {} + { + assert_invariant(); + } /*! @brief create an object (explicit) @@ -1041,7 +1111,9 @@ */ basic_json(const object_t& val) : m_type(value_t::object), m_value(val) - {} + { + assert_invariant(); + } /*! @brief create an object (implicit) @@ -1069,17 +1141,16 @@ @since version 1.0.0 */ - template ::value and - std::is_constructible::value, int>::type - = 0> + template::value and + std::is_constructible::value, int>::type = 0> basic_json(const CompatibleObjectType& val) : m_type(value_t::object) { using std::begin; using std::end; m_value.object = create(begin(val), end(val)); + assert_invariant(); } /*! @@ -1103,7 +1174,9 @@ */ basic_json(const array_t& val) : m_type(value_t::array), m_value(val) - {} + { + assert_invariant(); + } /*! @brief create an array (implicit) @@ -1131,22 +1204,21 @@ @since version 1.0.0 */ - template ::value and - not std::is_same::value and - not std::is_same::value and - not std::is_same::value and - not std::is_same::value and - not std::is_same::value and - std::is_constructible::value, int>::type - = 0> + template::value and + not std::is_same::value and + not std::is_same::value and + not std::is_same::value and + not std::is_same::value and + not std::is_same::value and + std::is_constructible::value, int>::type = 0> basic_json(const CompatibleArrayType& val) : m_type(value_t::array) { using std::begin; using std::end; m_value.array = create(begin(val), end(val)); + assert_invariant(); } /*! @@ -1172,7 +1244,9 @@ */ basic_json(const string_t& val) : m_type(value_t::string), m_value(val) - {} + { + assert_invariant(); + } /*! @brief create a string (explicit) @@ -1196,7 +1270,9 @@ */ basic_json(const typename string_t::value_type* val) : basic_json(string_t(val)) - {} + { + assert_invariant(); + } /*! @brief create a string (implicit) @@ -1221,13 +1297,13 @@ @since version 1.0.0 */ - template ::value, int>::type - = 0> + template::value, int>::type = 0> basic_json(const CompatibleStringType& val) : basic_json(string_t(val)) - {} + { + assert_invariant(); + } /*! @brief create a boolean (explicit) @@ -1245,7 +1321,9 @@ */ basic_json(boolean_t val) noexcept : m_type(value_t::boolean), m_value(val) - {} + { + assert_invariant(); + } /*! @brief create an integer number (explicit) @@ -1270,15 +1348,14 @@ @since version 1.0.0 */ - template::value) - and std::is_same::value - , int>::type - = 0> + template::value) and + std::is_same::value, int>::type = 0> basic_json(const number_integer_t val) noexcept : m_type(value_t::number_integer), m_value(val) - {} + { + assert_invariant(); + } /*! @brief create an integer number from an enum type (explicit) @@ -1308,7 +1385,9 @@ basic_json(const int val) noexcept : m_type(value_t::number_integer), m_value(static_cast(val)) - {} + { + assert_invariant(); + } /*! @brief create an integer number (implicit) @@ -1335,25 +1414,25 @@ @since version 1.0.0 */ - template::value and std::numeric_limits::is_integer and std::numeric_limits::is_signed, - CompatibleNumberIntegerType>::type - = 0> + CompatibleNumberIntegerType>::type = 0> basic_json(const CompatibleNumberIntegerType val) noexcept : m_type(value_t::number_integer), m_value(static_cast(val)) - {} + { + assert_invariant(); + } /*! @brief create an unsigned integer number (explicit) Create an unsigned integer number JSON value with a given content. - @tparam T helper type to compare number_unsigned_t and unsigned int - (not visible in) the interface. + @tparam T helper type to compare number_unsigned_t and unsigned int (not + visible in) the interface. @param[in] val an integer to create a JSON number from @@ -1364,15 +1443,14 @@ @since version 2.0.0 */ - template::value) - and std::is_same::value - , int>::type - = 0> + template::value) and + std::is_same::value, int>::type = 0> basic_json(const number_unsigned_t val) noexcept : m_type(value_t::number_unsigned), m_value(val) - {} + { + assert_invariant(); + } /*! @brief create an unsigned number (implicit) @@ -1394,17 +1472,17 @@ @since version 2.0.0 */ - template ::value and - std::numeric_limits::is_integer and - not std::numeric_limits::is_signed, - CompatibleNumberUnsignedType>::type - = 0> + template::value and + std::numeric_limits::is_integer and + not std::numeric_limits::is_signed, + CompatibleNumberUnsignedType>::type = 0> basic_json(const CompatibleNumberUnsignedType val) noexcept : m_type(value_t::number_unsigned), m_value(static_cast(val)) - {} + { + assert_invariant(); + } /*! @brief create a floating-point number (explicit) @@ -1417,8 +1495,8 @@ disallows NaN values: > Numeric values that cannot be represented in the grammar below (such as > Infinity and NaN) are not permitted. - In case the parameter @a val is not a number, a JSON null value is - created instead. + In case the parameter @a val is not a number, a JSON null value is created + instead. @complexity Constant. @@ -1439,6 +1517,8 @@ m_type = value_t::null; m_value = json_value(); } + + assert_invariant(); } /*! @@ -1472,14 +1552,14 @@ @since version 1.0.0 */ - template::value and - std::is_floating_point::value>::type - > + std::is_floating_point::value>::type> basic_json(const CompatibleNumberFloatType val) noexcept : basic_json(number_float_t(val)) - {} + { + assert_invariant(); + } /*! @brief create a container (array or object) from an initializer list @@ -1491,21 +1571,21 @@ 1. If the list is empty, an empty JSON object value `{}` is created. 2. If the list consists of pairs whose first element is a string, a JSON - object value is created where the first elements of the pairs are treated - as keys and the second elements are as values. + object value is created where the first elements of the pairs are + treated as keys and the second elements are as values. 3. In all other cases, an array is created. The rules aim to create the best fit between a C++ initializer list and JSON values. The rationale is as follows: 1. The empty initializer list is written as `{}` which is exactly an empty - JSON object. + JSON object. 2. C++ has now way of describing mapped types other than to list a list of - pairs. As JSON requires that keys must be of type string, rule 2 is the - weakest constraint one can pose on initializer lists to interpret them as - an object. + pairs. As JSON requires that keys must be of type string, rule 2 is the + weakest constraint one can pose on initializer lists to interpret them + as an object. 3. In all other cases, the initializer list could not be interpreted as - JSON object type, so interpreting it as JSON array type is safe. + JSON object type, so interpreting it as JSON array type is safe. With the rules described above, the following JSON values cannot be expressed by an initializer list: @@ -1554,22 +1634,13 @@ bool type_deduction = true, value_t manual_type = value_t::array) { - // the initializer list could describe an object - bool is_an_object = true; - // check if each element is an array with two elements whose first // element is a string - for (const auto& element : init) - { - if (not element.is_array() or element.size() != 2 - or not element[0].is_string()) - { - // we found an element that makes it impossible to use the - // initializer list as object - is_an_object = false; - break; - } - } + bool is_an_object = std::all_of(init.begin(), init.end(), + [](const basic_json & element) + { + return element.is_array() and element.size() == 2 and element[0].is_string(); + }); // adjust type if type deduction is not wanted if (not type_deduction) @@ -1593,12 +1664,10 @@ m_type = value_t::object; m_value = value_t::object; - assert(m_value.object != nullptr); - - for (auto& element : init) + std::for_each(init.begin(), init.end(), [this](const basic_json & element) { m_value.object->emplace(*(element[0].m_value.string), element[1]); - } + }); } else { @@ -1606,6 +1675,8 @@ m_type = value_t::array; m_value.array = create(init); } + + assert_invariant(); } /*! @@ -1710,6 +1781,7 @@ : m_type(value_t::array) { m_value.array = create(cnt, val); + assert_invariant(); } /*! @@ -1730,6 +1802,9 @@ @param[in] first begin of the range to copy from (included) @param[in] last end of the range to copy from (excluded) + @pre Iterators @a first and @a last must be initialized. **This + precondition is enforced with an assertion.** + @throw std::domain_error if iterators are not compatible; that is, do not belong to the same JSON value; example: `"iterators are not compatible"` @throw std::out_of_range if iterators are for a primitive type (number, @@ -1746,20 +1821,23 @@ @since version 1.0.0 */ - template ::value or - std::is_same::value - , int>::type - = 0> - basic_json(InputIT first, InputIT last) : m_type(first.m_object->m_type) - { + template::value or + std::is_same::value, int>::type = 0> + basic_json(InputIT first, InputIT last) + { + assert(first.m_object != nullptr); + assert(last.m_object != nullptr); + // make sure iterator fits the current value if (first.m_object != last.m_object) { throw std::domain_error("iterators are not compatible"); } + // copy type from first iterator + m_type = first.m_object->m_type; + // check if iterator range is complete for primitive values switch (m_type) { @@ -1786,35 +1864,30 @@ { case value_t::number_integer: { - assert(first.m_object != nullptr); m_value.number_integer = first.m_object->m_value.number_integer; break; } case value_t::number_unsigned: { - assert(first.m_object != nullptr); m_value.number_unsigned = first.m_object->m_value.number_unsigned; break; } case value_t::number_float: { - assert(first.m_object != nullptr); m_value.number_float = first.m_object->m_value.number_float; break; } case value_t::boolean: { - assert(first.m_object != nullptr); m_value.boolean = first.m_object->m_value.boolean; break; } case value_t::string: { - assert(first.m_object != nullptr); m_value = *first.m_object->m_value.string; break; } @@ -1833,10 +1906,11 @@ default: { - assert(first.m_object != nullptr); throw std::domain_error("cannot use construct with iterators from " + first.m_object->type_name()); } } + + assert_invariant(); } /*! @@ -1853,15 +1927,25 @@ @note A UTF-8 byte order mark is silently ignored. + @deprecated This constructor is deprecated and will be removed in version + 3.0.0 to unify the interface of the library. Deserialization will be + done by stream operators or by calling one of the `parse` functions, + e.g. @ref parse(std::istream&, const parser_callback_t). That is, calls + like `json j(i);` for an input stream @a i need to be replaced by + `json j = json::parse(i);`. See the example below. + @liveexample{The example below demonstrates constructing a JSON value from a `std::stringstream` with and without callback function.,basic_json__istream} - @since version 2.0.0 - */ - explicit basic_json(std::istream& i, parser_callback_t cb = nullptr) + @since version 2.0.0, deprecated in version 2.0.3, to be removed in + version 3.0.0 + */ + JSON_DEPRECATED + explicit basic_json(std::istream& i, const parser_callback_t cb = nullptr) { *this = parser(i, cb).parse(); + assert_invariant(); } /////////////////////////////////////// @@ -1893,25 +1977,25 @@ basic_json(const basic_json& other) : m_type(other.m_type) { + // check of passed value is valid + other.assert_invariant(); + switch (m_type) { case value_t::object: { - assert(other.m_value.object != nullptr); m_value = *other.m_value.object; break; } case value_t::array: { - assert(other.m_value.array != nullptr); m_value = *other.m_value.array; break; } case value_t::string: { - assert(other.m_value.string != nullptr); m_value = *other.m_value.string; break; } @@ -1945,6 +2029,8 @@ break; } } + + assert_invariant(); } /*! @@ -1969,9 +2055,14 @@ : m_type(std::move(other.m_type)), m_value(std::move(other.m_value)) { + // check that passed value is valid + other.assert_invariant(); + // invalidate payload other.m_type = value_t::null; other.m_value = {}; + + assert_invariant(); } /*! @@ -2004,9 +2095,14 @@ std::is_nothrow_move_assignable::value ) { + // check that passed value is valid + other.assert_invariant(); + using std::swap; swap(m_type, other.m_type); swap(m_value, other.m_value); + + assert_invariant(); return *this; } @@ -2027,6 +2123,8 @@ */ ~basic_json() { + assert_invariant(); + switch (m_type) { case value_t::object: @@ -2069,19 +2167,20 @@ /////////////////////// /// @name object inspection + /// Functions to inspect the type of a JSON value. /// @{ /*! @brief serialization Serialization function for JSON values. The function tries to mimic - Python's @p json.dumps() function, and currently supports its @p indent + Python's `json.dumps()` function, and currently supports its @a indent parameter. - @param[in] indent if indent is nonnegative, then array elements and object + @param[in] indent If indent is nonnegative, then array elements and object members will be pretty-printed with that indent level. An indent level of - 0 will only insert newlines. -1 (the default) selects the most compact - representation + `0` will only insert newlines. `-1` (the default) selects the most compact + representation. @return string containing the serialization of the JSON value @@ -2098,7 +2197,14 @@ { std::stringstream ss; // fix locale problems - ss.imbue(std::locale(std::locale(), new DecimalSeparator)); + const static std::locale loc(std::locale(), new DecimalSeparator); + ss.imbue(loc); + + // 6, 15 or 16 digits of precision allows round-trip IEEE 754 + // string->float->string, string->double->string or string->long + // double->string; to be safe, we read this value from + // std::numeric_limits::digits10 + ss.precision(std::numeric_limits::digits10); if (indent >= 0) { @@ -2475,16 +2581,13 @@ ////////////////// /// get an object (explicit) - template ::value and - std::is_convertible::value - , int>::type = 0> + template::value and + std::is_convertible::value, int>::type = 0> T get_impl(T*) const { if (is_object()) { - assert(m_value.object != nullptr); return T(m_value.object->begin(), m_value.object->end()); } else @@ -2498,7 +2601,6 @@ { if (is_object()) { - assert(m_value.object != nullptr); return *(m_value.object); } else @@ -2508,20 +2610,17 @@ } /// get an array (explicit) - template ::value and - not std::is_same::value and - not std::is_arithmetic::value and - not std::is_convertible::value and - not has_mapped_type::value - , int>::type = 0> + template::value and + not std::is_same::value and + not std::is_arithmetic::value and + not std::is_convertible::value and + not has_mapped_type::value, int>::type = 0> T get_impl(T*) const { if (is_array()) { T to_vector; - assert(m_value.array != nullptr); std::transform(m_value.array->begin(), m_value.array->end(), std::inserter(to_vector, to_vector.end()), [](basic_json i) { @@ -2536,17 +2635,14 @@ } /// get an array (explicit) - template ::value and - not std::is_same::value - , int>::type = 0> + template::value and + not std::is_same::value, int>::type = 0> std::vector get_impl(std::vector*) const { if (is_array()) { std::vector to_vector; - assert(m_value.array != nullptr); to_vector.reserve(m_value.array->size()); std::transform(m_value.array->begin(), m_value.array->end(), std::inserter(to_vector, to_vector.end()), [](basic_json i) @@ -2562,16 +2658,13 @@ } /// get an array (explicit) - template ::value and - not has_mapped_type::value - , int>::type = 0> + template::value and + not has_mapped_type::value, int>::type = 0> T get_impl(T*) const { if (is_array()) { - assert(m_value.array != nullptr); return T(m_value.array->begin(), m_value.array->end()); } else @@ -2585,7 +2678,6 @@ { if (is_array()) { - assert(m_value.array != nullptr); return *(m_value.array); } else @@ -2595,15 +2687,12 @@ } /// get a string (explicit) - template ::value - , int>::type = 0> + template::value, int>::type = 0> T get_impl(T*) const { if (is_string()) { - assert(m_value.string != nullptr); return *m_value.string; } else @@ -2613,10 +2702,8 @@ } /// get a number (explicit) - template::value - , int>::type = 0> + template::value, int>::type = 0> T get_impl(T*) const { switch (m_type) @@ -2749,8 +2836,10 @@ template static ReferenceType get_ref_impl(ThisType& obj) { + // helper type + using PointerType = typename std::add_pointer::type; + // delegate the call to get_ptr<>() - using PointerType = typename std::add_pointer::type; auto ptr = obj.template get_ptr(); if (ptr != nullptr) @@ -2767,6 +2856,7 @@ public: /// @name value access + /// Direct access to the stored value of a JSON value. /// @{ /*! @@ -2802,10 +2892,8 @@ @since version 1.0.0 */ - template::value - , int>::type = 0> + template::value, int>::type = 0> ValueType get() const { return get_impl(static_cast(nullptr)); @@ -2817,7 +2905,8 @@ Explicit pointer access to the internally stored JSON value. No copies are made. - @warning The pointer becomes invalid if the underlying JSON object changes. + @warning The pointer becomes invalid if the underlying JSON object + changes. @tparam PointerType pointer type; must be a pointer to @ref array_t, @ref object_t, @ref string_t, @ref boolean_t, @ref number_integer_t, @@ -2837,10 +2926,8 @@ @since version 1.0.0 */ - template::value - , int>::type = 0> + template::value, int>::type = 0> PointerType get() noexcept { // delegate the call to get_ptr @@ -2851,10 +2938,8 @@ @brief get a pointer value (explicit) @copydoc get() */ - template::value - , int>::type = 0> + template::value, int>::type = 0> constexpr const PointerType get() const noexcept { // delegate the call to get_ptr @@ -2872,7 +2957,8 @@ @tparam PointerType pointer type; must be a pointer to @ref array_t, @ref object_t, @ref string_t, @ref boolean_t, @ref number_integer_t, - @ref number_unsigned_t, or @ref number_float_t. + @ref number_unsigned_t, or @ref number_float_t. Enforced by a static + assertion. @return pointer to the internally stored JSON value if the requested pointer type @a PointerType fits to the JSON value; `nullptr` otherwise @@ -2886,12 +2972,25 @@ @since version 1.0.0 */ - template::value - , int>::type = 0> + template::value, int>::type = 0> PointerType get_ptr() noexcept { + // get the type of the PointerType (remove pointer and const) + using pointee_t = typename std::remove_const::type>::type>::type; + // make sure the type matches the allowed types + static_assert( + std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + , "incompatible pointer type"); + // delegate the call to get_impl_ptr<>() return get_impl_ptr(static_cast(nullptr)); } @@ -2900,13 +2999,26 @@ @brief get a pointer value (implicit) @copydoc get_ptr() */ - template::value - and std::is_const::type>::value - , int>::type = 0> + template::value and + std::is_const::type>::value, int>::type = 0> constexpr const PointerType get_ptr() const noexcept { + // get the type of the PointerType (remove pointer and const) + using pointee_t = typename std::remove_const::type>::type>::type; + // make sure the type matches the allowed types + static_assert( + std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + or std::is_same::value + , "incompatible pointer type"); + // delegate the call to get_impl_ptr<>() const return get_impl_ptr(static_cast(nullptr)); } @@ -2922,7 +3034,7 @@ @tparam ReferenceType reference type; must be a reference to @ref array_t, @ref object_t, @ref string_t, @ref boolean_t, @ref number_integer_t, or - @ref number_float_t. + @ref number_float_t. Enforced by static assertion. @return reference to the internally stored JSON value if the requested reference type @a ReferenceType fits to the JSON value; throws @@ -2937,10 +3049,8 @@ @since version 1.1.0 */ - template::value - , int>::type = 0> + template::value, int>::type = 0> ReferenceType get_ref() { // delegate call to get_ref_impl @@ -2951,11 +3061,9 @@ @brief get a reference value (implicit) @copydoc get_ref() */ - template::value - and std::is_const::type>::value - , int>::type = 0> + template::value and + std::is_const::type>::value, int>::type = 0> ReferenceType get_ref() const { // delegate call to get_ref_impl @@ -2990,10 +3098,9 @@ @since version 1.0.0 */ - template < typename ValueType, typename - std::enable_if < - not std::is_pointer::value - and not std::is_same::value + template < typename ValueType, typename std::enable_if < + not std::is_pointer::value and + not std::is_same::value #ifndef _MSC_VER // Fix for issue #167 operator<< abiguity under VS2015 and not std::is_same>::value #endif @@ -3012,6 +3119,7 @@ //////////////////// /// @name element access + /// Access to the JSON value. /// @{ /*! @@ -3043,7 +3151,6 @@ { try { - assert(m_value.array != nullptr); return m_value.array->at(idx); } catch (std::out_of_range&) @@ -3087,7 +3194,6 @@ { try { - assert(m_value.array != nullptr); return m_value.array->at(idx); } catch (std::out_of_range&) @@ -3135,7 +3241,6 @@ { try { - assert(m_value.object != nullptr); return m_value.object->at(key); } catch (std::out_of_range&) @@ -3183,7 +3288,6 @@ { try { - assert(m_value.object != nullptr); return m_value.object->at(key); } catch (std::out_of_range&) @@ -3230,16 +3334,18 @@ { m_type = value_t::array; m_value.array = create(); + assert_invariant(); } // operator[] only works for arrays if (is_array()) { - // fill up array with null values until given idx is reached - assert(m_value.array != nullptr); - for (size_t i = m_value.array->size(); i <= idx; ++i) - { - m_value.array->push_back(basic_json()); + // fill up array with null values if given idx is outside range + if (idx >= m_value.array->size()) + { + m_value.array->insert(m_value.array->end(), + idx - m_value.array->size() + 1, + basic_json()); } return m_value.array->operator[](idx); @@ -3274,7 +3380,6 @@ // const operator[] only works for arrays if (is_array()) { - assert(m_value.array != nullptr); return m_value.array->operator[](idx); } else @@ -3317,12 +3422,12 @@ { m_type = value_t::object; m_value.object = create(); + assert_invariant(); } // operator[] only works for objects if (is_object()) { - assert(m_value.object != nullptr); return m_value.object->operator[](key); } else @@ -3344,6 +3449,9 @@ @return const reference to the element at key @a key + @pre The element with key @a key must exist. **This precondition is + enforced with an assertion.** + @throw std::domain_error if JSON is not an object; example: `"cannot use operator[] with null"` @@ -3363,7 +3471,6 @@ // const operator[] only works for objects if (is_object()) { - assert(m_value.object != nullptr); assert(m_value.object->find(key) != m_value.object->end()); return m_value.object->find(key)->second; } @@ -3476,12 +3583,12 @@ { m_type = value_t::object; m_value = value_t::object; + assert_invariant(); } // at only works for objects if (is_object()) { - assert(m_value.object != nullptr); return m_value.object->operator[](key); } else @@ -3503,6 +3610,9 @@ @return const reference to the element at key @a key + @pre The element with key @a key must exist. **This precondition is + enforced with an assertion.** + @throw std::domain_error if JSON is not an object; example: `"cannot use operator[] with null"` @@ -3523,7 +3633,6 @@ // at only works for objects if (is_object()) { - assert(m_value.object != nullptr); assert(m_value.object->find(key) != m_value.object->end()); return m_value.object->find(key)->second; } @@ -3536,8 +3645,8 @@ /*! @brief access specified object element with default value - Returns either a copy of an object's element at the specified key @a key or - a given default value if no element with key @a key exists. + Returns either a copy of an object's element at the specified key @a key + or a given default value if no element with key @a key exists. The function is basically equivalent to executing @code {.cpp} @@ -3581,10 +3690,8 @@ @since version 1.0.0 */ - template ::value - , int>::type = 0> + template::value, int>::type = 0> ValueType value(const typename object_t::key_type& key, ValueType default_value) const { // at only works for objects @@ -3609,7 +3716,7 @@ /*! @brief overload for a default value of type const char* - @copydoc basic_json::value() + @copydoc basic_json::value(const typename object_t::key_type&, ValueType) const */ string_t value(const typename object_t::key_type& key, const char* default_value) const { @@ -3617,6 +3724,79 @@ } /*! + @brief access specified object element via JSON Pointer with default value + + Returns either a copy of an object's element at the specified key @a key + or a given default value if no element with key @a key exists. + + The function is basically equivalent to executing + @code {.cpp} + try { + return at(ptr); + } catch(std::out_of_range) { + return default_value; + } + @endcode + + @note Unlike @ref at(const json_pointer&), this function does not throw + if the given key @a key was not found. + + @param[in] ptr a JSON pointer to the element to access + @param[in] default_value the value to return if @a ptr found no value + + @tparam ValueType type compatible to JSON values, for instance `int` for + JSON integer numbers, `bool` for JSON booleans, or `std::vector` types for + JSON arrays. Note the type of the expected value at @a key and the default + value @a default_value must be compatible. + + @return copy of the element at key @a key or @a default_value if @a key + is not found + + @throw std::domain_error if JSON is not an object; example: `"cannot use + value() with null"` + + @complexity Logarithmic in the size of the container. + + @liveexample{The example below shows how object elements can be queried + with a default value.,basic_json__value_ptr} + + @sa @ref operator[](const json_pointer&) for unchecked access by reference + + @since version 2.0.2 + */ + template::value, int>::type = 0> + ValueType value(const json_pointer& ptr, ValueType default_value) const + { + // at only works for objects + if (is_object()) + { + // if pointer resolves a value, return it or use default value + try + { + return ptr.get_checked(this); + } + catch (std::out_of_range&) + { + return default_value; + } + } + else + { + throw std::domain_error("cannot use value() with " + type_name()); + } + } + + /*! + @brief overload for a default value of type const char* + @copydoc basic_json::value(const json_pointer&, ValueType) const + */ + string_t value(const json_pointer& ptr, const char* default_value) const + { + return value(ptr, string_t(default_value)); + } + + /*! @brief access the first element Returns a reference to the first element in the container. For a JSON @@ -3629,7 +3809,8 @@ @complexity Constant. @pre The JSON value must not be `null` (would throw `std::out_of_range`) - or an empty array or object (undefined behavior, guarded by assertions). + or an empty array or object (undefined behavior, **guarded by + assertions**). @post The JSON value remains unchanged. @throw std::out_of_range when called on `null` value @@ -3671,7 +3852,8 @@ @complexity Constant. @pre The JSON value must not be `null` (would throw `std::out_of_range`) - or an empty array or object (undefined behavior, guarded by assertions). + or an empty array or object (undefined behavior, **guarded by + assertions**). @post The JSON value remains unchanged. @throw std::out_of_range when called on `null` value. @@ -3713,7 +3895,7 @@ @return Iterator following the last removed element. If the iterator @a pos refers to the last element, the `end()` iterator is returned. - @tparam InteratorType an @ref iterator or @ref const_iterator + @tparam IteratorType an @ref iterator or @ref const_iterator @post Invalidates iterators and references at or after the point of the erase, including the `end()` iterator. @@ -3735,7 +3917,7 @@ @liveexample{The example shows the result of `erase()` for different JSON types.,erase__IteratorType} - @sa @ref erase(InteratorType, InteratorType) -- removes the elements in + @sa @ref erase(IteratorType, IteratorType) -- removes the elements in the given range @sa @ref erase(const typename object_t::key_type&) -- removes the element from an object at the given key @@ -3744,13 +3926,11 @@ @since version 1.0.0 */ - template ::value or - std::is_same::value - , int>::type - = 0> - InteratorType erase(InteratorType pos) + template::value or + std::is_same::value, int>::type + = 0> + IteratorType erase(IteratorType pos) { // make sure iterator fits the current value if (this != pos.m_object) @@ -3758,7 +3938,7 @@ throw std::domain_error("iterator does not fit current value"); } - InteratorType result = end(); + IteratorType result = end(); switch (m_type) { @@ -3775,24 +3955,25 @@ if (is_string()) { - delete m_value.string; + AllocatorType alloc; + alloc.destroy(m_value.string); + alloc.deallocate(m_value.string, 1); m_value.string = nullptr; } m_type = value_t::null; + assert_invariant(); break; } case value_t::object: { - assert(m_value.object != nullptr); result.m_it.object_iterator = m_value.object->erase(pos.m_it.object_iterator); break; } case value_t::array: { - assert(m_value.array != nullptr); result.m_it.array_iterator = m_value.array->erase(pos.m_it.array_iterator); break; } @@ -3821,7 +4002,7 @@ @return Iterator following the last removed element. If the iterator @a second refers to the last element, the `end()` iterator is returned. - @tparam InteratorType an @ref iterator or @ref const_iterator + @tparam IteratorType an @ref iterator or @ref const_iterator @post Invalidates iterators and references at or after the point of the erase, including the `end()` iterator. @@ -3844,7 +4025,7 @@ @liveexample{The example shows the result of `erase()` for different JSON types.,erase__IteratorType_IteratorType} - @sa @ref erase(InteratorType) -- removes the element at a given position + @sa @ref erase(IteratorType) -- removes the element at a given position @sa @ref erase(const typename object_t::key_type&) -- removes the element from an object at the given key @sa @ref erase(const size_type) -- removes the element from an array at @@ -3852,13 +4033,11 @@ @since version 1.0.0 */ - template ::value or - std::is_same::value - , int>::type - = 0> - InteratorType erase(InteratorType first, InteratorType last) + template::value or + std::is_same::value, int>::type + = 0> + IteratorType erase(IteratorType first, IteratorType last) { // make sure iterator fits the current value if (this != first.m_object or this != last.m_object) @@ -3866,7 +4045,7 @@ throw std::domain_error("iterators do not fit current value"); } - InteratorType result = end(); + IteratorType result = end(); switch (m_type) { @@ -3883,17 +4062,19 @@ if (is_string()) { - delete m_value.string; + AllocatorType alloc; + alloc.destroy(m_value.string); + alloc.deallocate(m_value.string, 1); m_value.string = nullptr; } m_type = value_t::null; + assert_invariant(); break; } case value_t::object: { - assert(m_value.object != nullptr); result.m_it.object_iterator = m_value.object->erase(first.m_it.object_iterator, last.m_it.object_iterator); break; @@ -3901,7 +4082,6 @@ case value_t::array: { - assert(m_value.array != nullptr); result.m_it.array_iterator = m_value.array->erase(first.m_it.array_iterator, last.m_it.array_iterator); break; @@ -3937,8 +4117,8 @@ @liveexample{The example shows the effect of `erase()`.,erase__key_type} - @sa @ref erase(InteratorType) -- removes the element at a given position - @sa @ref erase(InteratorType, InteratorType) -- removes the elements in + @sa @ref erase(IteratorType) -- removes the element at a given position + @sa @ref erase(IteratorType, IteratorType) -- removes the elements in the given range @sa @ref erase(const size_type) -- removes the element from an array at the given index @@ -3950,7 +4130,6 @@ // this erase only works for objects if (is_object()) { - assert(m_value.object != nullptr); return m_value.object->erase(key); } else @@ -3975,8 +4154,8 @@ @liveexample{The example shows the effect of `erase()`.,erase__size_type} - @sa @ref erase(InteratorType) -- removes the element at a given position - @sa @ref erase(InteratorType, InteratorType) -- removes the elements in + @sa @ref erase(IteratorType) -- removes the element at a given position + @sa @ref erase(IteratorType, IteratorType) -- removes the elements in the given range @sa @ref erase(const typename object_t::key_type&) -- removes the element from an object at the given key @@ -3993,7 +4172,6 @@ throw std::out_of_range("array index " + std::to_string(idx) + " is out of range"); } - assert(m_value.array != nullptr); m_value.array->erase(m_value.array->begin() + static_cast(idx)); } else @@ -4036,7 +4214,6 @@ if (is_object()) { - assert(m_value.object != nullptr); result.m_it.object_iterator = m_value.object->find(key); } @@ -4053,7 +4230,6 @@ if (is_object()) { - assert(m_value.object != nullptr); result.m_it.object_iterator = m_value.object->find(key); } @@ -4081,7 +4257,6 @@ size_type count(typename object_t::key_type key) const { // return 0 for all nonobject types - assert(not is_object() or m_value.object != nullptr); return is_object() ? m_value.object->count(key) : 0; } @@ -4423,6 +4598,10 @@ object | result of function `object_t::empty()` array | result of function `array_t::empty()` + @note This function does not return whether a string stored as JSON value + is empty - it returns whether the JSON container itself is empty which is + false in the case of a string. + @complexity Constant, as long as @ref array_t and @ref object_t satisfy the Container concept; that is, their `empty()` functions have constant complexity. @@ -4452,13 +4631,13 @@ case value_t::array: { - assert(m_value.array != nullptr); + // delegate call to array_t::empty() return m_value.array->empty(); } case value_t::object: { - assert(m_value.object != nullptr); + // delegate call to object_t::empty() return m_value.object->empty(); } @@ -4486,6 +4665,10 @@ object | result of function object_t::size() array | result of function array_t::size() + @note This function does not return the length of a string stored as JSON + value - it returns the number of elements in the JSON value which is 1 in + the case of a string. + @complexity Constant, as long as @ref array_t and @ref object_t satisfy the Container concept; that is, their size() functions have constant complexity. @@ -4516,13 +4699,13 @@ case value_t::array: { - assert(m_value.array != nullptr); + // delegate call to array_t::size() return m_value.array->size(); } case value_t::object: { - assert(m_value.object != nullptr); + // delegate call to object_t::size() return m_value.object->size(); } @@ -4576,13 +4759,13 @@ { case value_t::array: { - assert(m_value.array != nullptr); + // delegate call to array_t::max_size() return m_value.array->max_size(); } case value_t::object: { - assert(m_value.object != nullptr); + // delegate call to object_t::max_size() return m_value.object->max_size(); } @@ -4659,21 +4842,18 @@ case value_t::string: { - assert(m_value.string != nullptr); m_value.string->clear(); break; } case value_t::array: { - assert(m_value.array != nullptr); m_value.array->clear(); break; } case value_t::object: { - assert(m_value.object != nullptr); m_value.object->clear(); break; } @@ -4718,10 +4898,10 @@ { m_type = value_t::array; m_value = value_t::array; + assert_invariant(); } // add element to array (move semantics) - assert(m_value.array != nullptr); m_value.array->push_back(std::move(val)); // invalidate object val.m_type = value_t::null; @@ -4754,10 +4934,10 @@ { m_type = value_t::array; m_value = value_t::array; + assert_invariant(); } // add element to array - assert(m_value.array != nullptr); m_value.array->push_back(val); } @@ -4804,10 +4984,10 @@ { m_type = value_t::object; m_value = value_t::object; + assert_invariant(); } // add element to array - assert(m_value.object != nullptr); m_value.object->insert(val); } @@ -4904,7 +5084,6 @@ // insert to array and return iterator iterator result(this); - assert(m_value.array != nullptr); result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, val); return result; } @@ -4960,7 +5139,6 @@ // insert to array and return iterator iterator result(this); - assert(m_value.array != nullptr); result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, cnt, val); return result; } @@ -5027,7 +5205,6 @@ // insert to array and return iterator iterator result(this); - assert(m_value.array != nullptr); result.m_it.array_iterator = m_value.array->insert( pos.m_it.array_iterator, first.m_it.array_iterator, @@ -5075,7 +5252,6 @@ // insert to array and return iterator iterator result(this); - assert(m_value.array != nullptr); result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, ilist); return result; } @@ -5106,6 +5282,7 @@ { std::swap(m_type, other.m_type); std::swap(m_value, other.m_value); + assert_invariant(); } /*! @@ -5133,7 +5310,6 @@ // swap only works for arrays if (is_array()) { - assert(m_value.array != nullptr); std::swap(*(m_value.array), other); } else @@ -5167,7 +5343,6 @@ // swap only works for objects if (is_object()) { - assert(m_value.object != nullptr); std::swap(*(m_value.object), other); } else @@ -5201,7 +5376,6 @@ // swap only works for strings if (is_string()) { - assert(m_value.string != nullptr); std::swap(*(m_value.string), other); } else @@ -5288,14 +5462,10 @@ { case value_t::array: { - assert(lhs.m_value.array != nullptr); - assert(rhs.m_value.array != nullptr); return *lhs.m_value.array == *rhs.m_value.array; } case value_t::object: { - assert(lhs.m_value.object != nullptr); - assert(rhs.m_value.object != nullptr); return *lhs.m_value.object == *rhs.m_value.object; } case value_t::null: @@ -5304,8 +5474,6 @@ } case value_t::string: { - assert(lhs.m_value.string != nullptr); - assert(rhs.m_value.string != nullptr); return *lhs.m_value.string == *rhs.m_value.string; } case value_t::boolean: @@ -5478,14 +5646,10 @@ { case value_t::array: { - assert(lhs.m_value.array != nullptr); - assert(rhs.m_value.array != nullptr); return *lhs.m_value.array < *rhs.m_value.array; } case value_t::object: { - assert(lhs.m_value.object != nullptr); - assert(rhs.m_value.object != nullptr); return *lhs.m_value.object < *rhs.m_value.object; } case value_t::null: @@ -5494,8 +5658,6 @@ } case value_t::string: { - assert(lhs.m_value.string != nullptr); - assert(rhs.m_value.string != nullptr); return *lhs.m_value.string < *rhs.m_value.string; } case value_t::boolean: @@ -5637,6 +5799,10 @@ `std::setw(4)` on @a o sets the indentation level to `4` and the serialization result is the same as calling `dump(4)`. + @note During serializaion, the locale and the precision of the output + stream @a o are changed. The original values are restored when the + function returns. + @param[in,out] o stream to serialize to @param[in] j JSON value to serialize @@ -5657,14 +5823,23 @@ // reset width to 0 for subsequent calls to this stream o.width(0); + // fix locale problems - auto old_locale = o.imbue(std::locale(std::locale(), new DecimalSeparator)); + const auto old_locale = o.imbue(std::locale(std::locale(), new DecimalSeparator)); + // set precision + + // 6, 15 or 16 digits of precision allows round-trip IEEE 754 + // string->float->string, string->double->string or string->long + // double->string; to be safe, we read this value from + // std::numeric_limits::digits10 + const auto old_precision = o.precision(std::numeric_limits::digits10); // do the actual serialization j.dump(o, pretty_print, static_cast(indentation)); - // reset locale + // reset locale and precision o.imbue(old_locale); + o.precision(old_precision); return o; } @@ -5688,9 +5863,45 @@ /// @{ /*! - @brief deserialize from string - - @param[in] s string to read a serialized JSON value from + @brief deserialize from an array + + This function reads from an array of 1-byte values. + + @pre Each element of the container has a size of 1 byte. Violating this + precondition yields undefined behavior. **This precondition is enforced + with a static assertion.** + + @param[in] array array to read from + @param[in] cb a parser callback function of type @ref parser_callback_t + which is used to control the deserialization by filtering unwanted values + (optional) + + @return result of the deserialization + + @complexity Linear in the length of the input. The parser is a predictive + LL(1) parser. The complexity can be higher if the parser callback function + @a cb has a super-linear complexity. + + @note A UTF-8 byte order mark is silently ignored. + + @liveexample{The example below demonstrates the `parse()` function reading + from an array.,parse__array__parser_callback_t} + + @since version 2.0.3 + */ + template + static basic_json parse(T (&array)[N], + const parser_callback_t cb = nullptr) + { + // delegate the call to the iterator-range parse overload + return parse(std::begin(array), std::end(array), cb); + } + + /*! + @brief deserialize from string literal + + @tparam CharT character/literal type with size of 1 byte + @param[in] s string literal to read a serialized JSON value from @param[in] cb a parser callback function of type @ref parser_callback_t which is used to control the deserialization by filtering unwanted values (optional) @@ -5702,18 +5913,25 @@ @a cb has a super-linear complexity. @note A UTF-8 byte order mark is silently ignored. + @note String containers like `std::string` or @ref string_t can be parsed + with @ref parse(const ContiguousContainer&, const parser_callback_t) @liveexample{The example below demonstrates the `parse()` function with and without callback function.,parse__string__parser_callback_t} - @sa @ref parse(std::istream&, parser_callback_t) for a version that reads - from an input stream - - @since version 1.0.0 - */ - static basic_json parse(const string_t& s, parser_callback_t cb = nullptr) - { - return parser(s, cb).parse(); + @sa @ref parse(std::istream&, const parser_callback_t) for a version that + reads from an input stream + + @since version 1.0.0 (originally for @ref string_t) + */ + template::value and + std::is_integral::type>::value and + sizeof(typename std::remove_pointer::type) == 1, int>::type = 0> + static basic_json parse(const CharPT s, + const parser_callback_t cb = nullptr) + { + return parser(reinterpret_cast(s), cb).parse(); } /*! @@ -5735,22 +5953,148 @@ @liveexample{The example below demonstrates the `parse()` function with and without callback function.,parse__istream__parser_callback_t} - @sa @ref parse(const string_t&, parser_callback_t) for a version that - reads from a string - - @since version 1.0.0 - */ - static basic_json parse(std::istream& i, parser_callback_t cb = nullptr) + @sa @ref parse(const char*, const parser_callback_t) for a version + that reads from a string + + @since version 1.0.0 + */ + static basic_json parse(std::istream& i, + const parser_callback_t cb = nullptr) + { + return parser(i, cb).parse(); + } + + /*! + @copydoc parse(std::istream&, const parser_callback_t) + */ + static basic_json parse(std::istream&& i, + const parser_callback_t cb = nullptr) { return parser(i, cb).parse(); } /*! - @copydoc parse(std::istream&, parser_callback_t) - */ - static basic_json parse(std::istream&& i, parser_callback_t cb = nullptr) - { - return parser(i, cb).parse(); + @brief deserialize from an iterator range with contiguous storage + + This function reads from an iterator range of a container with contiguous + storage of 1-byte values. Compatible container types include + `std::vector`, `std::string`, `std::array`, `std::valarray`, and + `std::initializer_list`. Furthermore, C-style arrays can be used with + `std::begin()`/`std::end()`. User-defined containers can be used as long + as they implement random-access iterators and a contiguous storage. + + @pre The iterator range is contiguous. Violating this precondition yields + undefined behavior. **This precondition is enforced with an assertion.** + @pre Each element in the range has a size of 1 byte. Violating this + precondition yields undefined behavior. **This precondition is enforced + with a static assertion.** + + @warning There is no way to enforce all preconditions at compile-time. If + the function is called with noncompliant iterators and with + assertions switched off, the behavior is undefined and will most + likely yield segmentation violation. + + @tparam IteratorType iterator of container with contiguous storage + @param[in] first begin of the range to parse (included) + @param[in] last end of the range to parse (excluded) + @param[in] cb a parser callback function of type @ref parser_callback_t + which is used to control the deserialization by filtering unwanted values + (optional) + + @return result of the deserialization + + @complexity Linear in the length of the input. The parser is a predictive + LL(1) parser. The complexity can be higher if the parser callback function + @a cb has a super-linear complexity. + + @note A UTF-8 byte order mark is silently ignored. + + @liveexample{The example below demonstrates the `parse()` function reading + from an iterator range.,parse__iteratortype__parser_callback_t} + + @since version 2.0.3 + */ + template::iterator_category>::value, int>::type = 0> + static basic_json parse(IteratorType first, IteratorType last, + const parser_callback_t cb = nullptr) + { + // assertion to check that the iterator range is indeed contiguous, + // see http://stackoverflow.com/a/35008842/266378 for more discussion + assert(std::accumulate(first, last, std::make_pair(true, 0), + [&first](std::pair res, decltype(*first) val) + { + res.first &= (val == *(std::next(std::addressof(*first), res.second++))); + return res; + }).first); + + // assertion to check that each element is 1 byte long + static_assert(sizeof(typename std::iterator_traits::value_type) == 1, + "each element in the iterator range must have the size of 1 byte"); + + // if iterator range is empty, create a parser with an empty string + // to generate "unexpected EOF" error message + if (std::distance(first, last) <= 0) + { + return parser("").parse(); + } + + return parser(first, last, cb).parse(); + } + + /*! + @brief deserialize from a container with contiguous storage + + This function reads from a container with contiguous storage of 1-byte + values. Compatible container types include `std::vector`, `std::string`, + `std::array`, and `std::initializer_list`. User-defined containers can be + used as long as they implement random-access iterators and a contiguous + storage. + + @pre The container storage is contiguous. Violating this precondition + yields undefined behavior. **This precondition is enforced with an + assertion.** + @pre Each element of the container has a size of 1 byte. Violating this + precondition yields undefined behavior. **This precondition is enforced + with a static assertion.** + + @warning There is no way to enforce all preconditions at compile-time. If + the function is called with a noncompliant container and with + assertions switched off, the behavior is undefined and will most + likely yield segmentation violation. + + @tparam ContiguousContainer container type with contiguous storage + @param[in] c container to read from + @param[in] cb a parser callback function of type @ref parser_callback_t + which is used to control the deserialization by filtering unwanted values + (optional) + + @return result of the deserialization + + @complexity Linear in the length of the input. The parser is a predictive + LL(1) parser. The complexity can be higher if the parser callback function + @a cb has a super-linear complexity. + + @note A UTF-8 byte order mark is silently ignored. + + @liveexample{The example below demonstrates the `parse()` function reading + from a contiguous container.,parse__contiguouscontainer__parser_callback_t} + + @since version 2.0.3 + */ + template::value and + std::is_base_of< + std::random_access_iterator_tag, + typename std::iterator_traits()))>::iterator_category>::value + , int>::type = 0> + static basic_json parse(const ContiguousContainer& c, + const parser_callback_t cb = nullptr) + { + // delegate the call to the iterator-range parse overload + return parse(std::begin(c), std::end(c), cb); } /*! @@ -5771,8 +6115,8 @@ @liveexample{The example below shows how a JSON value is constructed by reading a serialization from a stream.,operator_deserialize} - @sa parse(std::istream&, parser_callback_t) for a variant with a parser - callback function to filter values while parsing + @sa parse(std::istream&, const parser_callback_t) for a variant with a + parser callback function to filter values while parsing @since version 1.0.0 */ @@ -5800,8 +6144,19 @@ // convenience functions // /////////////////////////// - /// return the type as string - string_t type_name() const noexcept + /*! + @brief return the type as string + + Returns the type name as string to be used in error messages - usually to + indicate that a function was called on a wrong JSON type. + + @return basically a string representation of a the @a m_type member + + @complexity Constant. + + @since version 1.0.0 + */ + std::string type_name() const { switch (m_type) { @@ -5832,9 +6187,8 @@ */ static std::size_t extra_space(const string_t& s) noexcept { - std::size_t result = 0; - - for (const auto& c : s) + return std::accumulate(s.begin(), s.end(), size_t{}, + [](size_t res, typename string_t::value_type c) { switch (c) { @@ -5847,8 +6201,7 @@ case '\t': { // from c (1 byte) to \x (2 bytes) - result += 1; - break; + return res + 1; } default: @@ -5856,14 +6209,15 @@ if (c >= 0x00 and c <= 0x1f) { // from c (1 byte) to \uxxxx (6 bytes) - result += 5; - } - break; - } - } - } - - return result; + return res + 5; + } + else + { + return res; + } + } + } + }); } /*! @@ -5957,16 +6311,15 @@ { // convert a number 0..15 to its hex representation // (0..f) - const auto hexify = [](const int v) -> char + static const char hexify[16] = { - return (v < 10) - ? ('0' + static_cast(v)) - : ('a' + static_cast((v - 10) & 0x1f)); + '0', '1', '2', '3', '4', '5', '6', '7', + '8', '9', 'a', 'b', 'c', 'd', 'e', 'f' }; // print character c as \uxxxx for (const char m : - { 'u', '0', '0', hexify(c >> 4), hexify(c & 0x0f) + { 'u', '0', '0', hexify[c >> 4], hexify[c & 0x0f] }) { result[++pos] = m; @@ -6016,8 +6369,6 @@ { case value_t::object: { - assert(m_value.object != nullptr); - if (m_value.object->empty()) { o << "{}"; @@ -6058,8 +6409,6 @@ case value_t::array: { - assert(m_value.array != nullptr); - if (m_value.array->empty()) { o << "[]"; @@ -6098,7 +6447,6 @@ case value_t::string: { - assert(m_value.string != nullptr); o << string_t("\"") << escape_string(*m_value.string) << "\""; return; } @@ -6130,13 +6478,7 @@ } else { - // Otherwise 6, 15 or 16 digits of precision allows - // round-trip IEEE 754 string->float->string, - // string->double->string or string->long - // double->string; to be safe, we read this value from - // std::numeric_limits::digits10 - o << std::setprecision(std::numeric_limits::digits10) - << m_value.number_float; + o << m_value.number_float; } return; } @@ -6353,6 +6695,12 @@ This class implements a const iterator for the @ref basic_json class. From this class, the @ref iterator class is derived. + @note An iterator is called *initialized* when a pointer to a JSON value + has been set (e.g., by a constructor or a copy assignment). If the + iterator is default-constructed, it is *uninitialized* and most + methods are undefined. **The library uses assertions to detect calls + on uninitialized iterators.** + @requirement The class satisfies the following concept requirements: - [RandomAccessIterator](http://en.cppreference.com/w/cpp/concept/RandomAccessIterator): The iterator that can be moved to point (forward and backward) to any @@ -6380,7 +6728,12 @@ /// default constructor const_iterator() = default; - /// constructor for a given JSON instance + /*! + @brief constructor for a given JSON instance + @param[in] object pointer to a JSON object for this iterator + @pre object != nullptr + @post The iterator is initialized; i.e. `m_object != nullptr`. + */ explicit const_iterator(pointer object) noexcept : m_object(object) { @@ -6408,40 +6761,53 @@ } } - /// copy constructor given a nonconst iterator + /*! + @brief copy constructor given a non-const iterator + @param[in] other iterator to copy from + @note It is not checked whether @a other is initialized. + */ explicit const_iterator(const iterator& other) noexcept : m_object(other.m_object) { - assert(m_object != nullptr); - - switch (m_object->m_type) - { - case basic_json::value_t::object: - { - m_it.object_iterator = other.m_it.object_iterator; - break; - } - - case basic_json::value_t::array: - { - m_it.array_iterator = other.m_it.array_iterator; - break; - } - - default: - { - m_it.primitive_iterator = other.m_it.primitive_iterator; - break; - } - } - } - - /// copy constructor + if (m_object != nullptr) + { + switch (m_object->m_type) + { + case basic_json::value_t::object: + { + m_it.object_iterator = other.m_it.object_iterator; + break; + } + + case basic_json::value_t::array: + { + m_it.array_iterator = other.m_it.array_iterator; + break; + } + + default: + { + m_it.primitive_iterator = other.m_it.primitive_iterator; + break; + } + } + } + } + + /*! + @brief copy constructor + @param[in] other iterator to copy from + @note It is not checked whether @a other is initialized. + */ const_iterator(const const_iterator& other) noexcept : m_object(other.m_object), m_it(other.m_it) {} - /// copy assignment + /*! + @brief copy assignment + @param[in,out] other iterator to copy from + @note It is not checked whether @a other is initialized. + */ const_iterator& operator=(const_iterator other) noexcept( std::is_nothrow_move_constructible::value and std::is_nothrow_move_assignable::value and @@ -6455,7 +6821,10 @@ } private: - /// set the iterator to the first value + /*! + @brief set the iterator to the first value + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ void set_begin() noexcept { assert(m_object != nullptr); @@ -6464,14 +6833,12 @@ { case basic_json::value_t::object: { - assert(m_object->m_value.object != nullptr); m_it.object_iterator = m_object->m_value.object->begin(); break; } case basic_json::value_t::array: { - assert(m_object->m_value.array != nullptr); m_it.array_iterator = m_object->m_value.array->begin(); break; } @@ -6491,7 +6858,10 @@ } } - /// set the iterator past the last value + /*! + @brief set the iterator past the last value + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ void set_end() noexcept { assert(m_object != nullptr); @@ -6500,14 +6870,12 @@ { case basic_json::value_t::object: { - assert(m_object->m_value.object != nullptr); m_it.object_iterator = m_object->m_value.object->end(); break; } case basic_json::value_t::array: { - assert(m_object->m_value.array != nullptr); m_it.array_iterator = m_object->m_value.array->end(); break; } @@ -6521,7 +6889,10 @@ } public: - /// return a reference to the value pointed to by the iterator + /*! + @brief return a reference to the value pointed to by the iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ reference operator*() const { assert(m_object != nullptr); @@ -6530,14 +6901,12 @@ { case basic_json::value_t::object: { - assert(m_object->m_value.object); assert(m_it.object_iterator != m_object->m_value.object->end()); return m_it.object_iterator->second; } case basic_json::value_t::array: { - assert(m_object->m_value.array); assert(m_it.array_iterator != m_object->m_value.array->end()); return *m_it.array_iterator; } @@ -6561,7 +6930,10 @@ } } - /// dereference the iterator + /*! + @brief dereference the iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ pointer operator->() const { assert(m_object != nullptr); @@ -6570,14 +6942,12 @@ { case basic_json::value_t::object: { - assert(m_object->m_value.object); assert(m_it.object_iterator != m_object->m_value.object->end()); return &(m_it.object_iterator->second); } case basic_json::value_t::array: { - assert(m_object->m_value.array); assert(m_it.array_iterator != m_object->m_value.array->end()); return &*m_it.array_iterator; } @@ -6596,7 +6966,10 @@ } } - /// post-increment (it++) + /*! + @brief post-increment (it++) + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator operator++(int) { auto result = *this; @@ -6604,7 +6977,10 @@ return result; } - /// pre-increment (++it) + /*! + @brief pre-increment (++it) + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator& operator++() { assert(m_object != nullptr); @@ -6613,13 +6989,13 @@ { case basic_json::value_t::object: { - ++m_it.object_iterator; + std::advance(m_it.object_iterator, 1); break; } case basic_json::value_t::array: { - ++m_it.array_iterator; + std::advance(m_it.array_iterator, 1); break; } @@ -6633,7 +7009,10 @@ return *this; } - /// post-decrement (it--) + /*! + @brief post-decrement (it--) + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator operator--(int) { auto result = *this; @@ -6641,7 +7020,10 @@ return result; } - /// pre-decrement (--it) + /*! + @brief pre-decrement (--it) + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator& operator--() { assert(m_object != nullptr); @@ -6650,13 +7032,13 @@ { case basic_json::value_t::object: { - --m_it.object_iterator; + std::advance(m_it.object_iterator, -1); break; } case basic_json::value_t::array: { - --m_it.array_iterator; + std::advance(m_it.array_iterator, -1); break; } @@ -6670,7 +7052,10 @@ return *this; } - /// comparison: equal + /*! + @brief comparison: equal + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ bool operator==(const const_iterator& other) const { // if objects are not the same, the comparison is undefined @@ -6700,13 +7085,19 @@ } } - /// comparison: not equal + /*! + @brief comparison: not equal + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ bool operator!=(const const_iterator& other) const { return not operator==(other); } - /// comparison: smaller + /*! + @brief comparison: smaller + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ bool operator<(const const_iterator& other) const { // if objects are not the same, the comparison is undefined @@ -6736,25 +7127,37 @@ } } - /// comparison: less than or equal + /*! + @brief comparison: less than or equal + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ bool operator<=(const const_iterator& other) const { return not other.operator < (*this); } - /// comparison: greater than + /*! + @brief comparison: greater than + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ bool operator>(const const_iterator& other) const { return not operator<=(other); } - /// comparison: greater than or equal + /*! + @brief comparison: greater than or equal + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ bool operator>=(const const_iterator& other) const { return not operator<(other); } - /// add to iterator + /*! + @brief add to iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator& operator+=(difference_type i) { assert(m_object != nullptr); @@ -6768,7 +7171,7 @@ case basic_json::value_t::array: { - m_it.array_iterator += i; + std::advance(m_it.array_iterator, i); break; } @@ -6782,13 +7185,19 @@ return *this; } - /// subtract from iterator + /*! + @brief subtract from iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator& operator-=(difference_type i) { return operator+=(-i); } - /// add to iterator + /*! + @brief add to iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator operator+(difference_type i) { auto result = *this; @@ -6796,7 +7205,10 @@ return result; } - /// subtract from iterator + /*! + @brief subtract from iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ const_iterator operator-(difference_type i) { auto result = *this; @@ -6804,7 +7216,10 @@ return result; } - /// return difference + /*! + @brief return difference + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ difference_type operator-(const const_iterator& other) const { assert(m_object != nullptr); @@ -6828,7 +7243,10 @@ } } - /// access to successor + /*! + @brief access to successor + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ reference operator[](difference_type n) const { assert(m_object != nullptr); @@ -6842,7 +7260,7 @@ case basic_json::value_t::array: { - return *(m_it.array_iterator + n); + return *std::next(m_it.array_iterator, n); } case basic_json::value_t::null: @@ -6864,7 +7282,10 @@ } } - /// return the key of an object iterator + /*! + @brief return the key of an object iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ typename object_t::key_type key() const { assert(m_object != nullptr); @@ -6879,7 +7300,10 @@ } } - /// return the value of an iterator + /*! + @brief return the value of an iterator + @pre The iterator is initialized; i.e. `m_object != nullptr`. + */ reference value() const { return operator*(); @@ -7177,54 +7601,55 @@ /// the char type to use in the lexer using lexer_char_t = unsigned char; - /// constructor with a given buffer - explicit lexer(const string_t& s) noexcept - : m_stream(nullptr), m_buffer(s) - { - m_content = reinterpret_cast(s.c_str()); + /// a lexer from a buffer with given length + lexer(const lexer_char_t* buff, const size_t len) noexcept + : m_content(buff) + { assert(m_content != nullptr); m_start = m_cursor = m_content; - m_limit = m_content + s.size(); - } - - /// constructor with a given stream - explicit lexer(std::istream* s) noexcept - : m_stream(s), m_buffer() - { - assert(m_stream != nullptr); - getline(*m_stream, m_buffer); - m_content = reinterpret_cast(m_buffer.c_str()); - assert(m_content != nullptr); - m_start = m_cursor = m_content; - m_limit = m_content + m_buffer.size(); - } - - /// default constructor - lexer() = default; - - // switch off unwanted functions + m_limit = m_content + len; + } + + /// a lexer from an input stream + explicit lexer(std::istream& s) + : m_stream(&s), m_line_buffer() + { + // fill buffer + fill_line_buffer(); + } + + // switch off unwanted functions (due to pointer members) + lexer() = delete; lexer(const lexer&) = delete; lexer operator=(const lexer&) = delete; /*! - @brief create a string from a Unicode code point + @brief create a string from one or two Unicode code points + + There are two cases: (1) @a codepoint1 is in the Basic Multilingual + Plane (U+0000 through U+FFFF) and @a codepoint2 is 0, or (2) + @a codepoint1 and @a codepoint2 are a UTF-16 surrogate pair to + represent a code point above U+FFFF. @param[in] codepoint1 the code point (can be high surrogate) @param[in] codepoint2 the code point (can be low surrogate or 0) - @return string representation of the code point + @return string representation of the code point; the length of the + result string is between 1 and 4 characters. @throw std::out_of_range if code point is > 0x10ffff; example: `"code points above 0x10FFFF are invalid"` @throw std::invalid_argument if the low surrogate is invalid; example: `""missing or wrong low surrogate""` + @complexity Constant. + @see */ static string_t to_unicode(const std::size_t codepoint1, const std::size_t codepoint2 = 0) { - // calculate the codepoint from the given code points + // calculate the code point from the given code points std::size_t codepoint = codepoint1; // check if codepoint1 is a high surrogate @@ -7286,7 +7711,7 @@ } /// return name of values of type token_type (only used for errors) - static std::string token_type_name(token_type t) + static std::string token_type_name(const token_type t) { switch (t) { @@ -7335,403 +7760,416 @@ function consists of a large block of code with `goto` jumps. @return the class of the next token read from the buffer + + @complexity Linear in the length of the input.\n + + Proposition: The loop below will always terminate for finite input.\n + + Proof (by contradiction): Assume a finite input. To loop forever, the + loop must never hit code with a `break` statement. The only code + snippets without a `break` statement are the continue statements for + whitespace and byte-order-marks. To loop forever, the input must be an + infinite sequence of whitespace or byte-order-marks. This contradicts + the assumption of finite input, q.e.d. */ - token_type scan() noexcept - { - // pointer for backtracking information - m_marker = nullptr; - - // remember the begin of the token - m_start = m_cursor; - assert(m_start != nullptr); - - - { - lexer_char_t yych; - unsigned int yyaccept = 0; - static const unsigned char yybm[] = - { - 0, 0, 0, 0, 0, 0, 0, 0, - 0, 32, 32, 0, 0, 32, 0, 0, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 160, 128, 0, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 0, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - 128, 128, 128, 128, 128, 128, 128, 128, - }; - if ((m_limit - m_cursor) < 5) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yybm[0 + yych] & 32) - { - goto basic_json_parser_6; - } - if (yych <= '\\') - { - if (yych <= '-') - { - if (yych <= '"') + token_type scan() + { + while (true) + { + // pointer for backtracking information + m_marker = nullptr; + + // remember the begin of the token + m_start = m_cursor; + assert(m_start != nullptr); + + + { + lexer_char_t yych; + unsigned int yyaccept = 0; + static const unsigned char yybm[] = + { + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 32, 32, 0, 0, 32, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, + 160, 128, 0, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 0, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + 128, 128, 128, 128, 128, 128, 128, 128, + }; + if ((m_limit - m_cursor) < 5) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yybm[0 + yych] & 32) + { + goto basic_json_parser_6; + } + if (yych <= '\\') + { + if (yych <= '-') { - if (yych <= 0x00) + if (yych <= '"') { - goto basic_json_parser_2; + if (yych <= 0x00) + { + goto basic_json_parser_2; + } + if (yych <= '!') + { + goto basic_json_parser_4; + } + goto basic_json_parser_9; } - if (yych <= '!') + else { - goto basic_json_parser_4; + if (yych <= '+') + { + goto basic_json_parser_4; + } + if (yych <= ',') + { + goto basic_json_parser_10; + } + goto basic_json_parser_12; } - goto basic_json_parser_9; } else { - if (yych <= '+') + if (yych <= '9') { + if (yych <= '/') + { + goto basic_json_parser_4; + } + if (yych <= '0') + { + goto basic_json_parser_13; + } + goto basic_json_parser_15; + } + else + { + if (yych <= ':') + { + goto basic_json_parser_17; + } + if (yych == '[') + { + goto basic_json_parser_19; + } goto basic_json_parser_4; } - if (yych <= ',') - { - goto basic_json_parser_10; - } - goto basic_json_parser_12; } } else { - if (yych <= '9') - { - if (yych <= '/') - { - goto basic_json_parser_4; - } - if (yych <= '0') - { - goto basic_json_parser_13; - } - goto basic_json_parser_15; - } - else - { - if (yych <= ':') - { - goto basic_json_parser_17; - } - if (yych == '[') - { - goto basic_json_parser_19; - } - goto basic_json_parser_4; - } - } - } - else - { - if (yych <= 't') - { - if (yych <= 'f') + if (yych <= 't') { - if (yych <= ']') - { - goto basic_json_parser_21; - } - if (yych <= 'e') + if (yych <= 'f') { - goto basic_json_parser_4; - } - goto basic_json_parser_23; - } - else - { - if (yych == 'n') - { - goto basic_json_parser_24; + if (yych <= ']') + { + goto basic_json_parser_21; + } + if (yych <= 'e') + { + goto basic_json_parser_4; + } + goto basic_json_parser_23; } - if (yych <= 's') + else { - goto basic_json_parser_4; + if (yych == 'n') + { + goto basic_json_parser_24; + } + if (yych <= 's') + { + goto basic_json_parser_4; + } + goto basic_json_parser_25; } - goto basic_json_parser_25; - } - } - else - { - if (yych <= '|') - { - if (yych == '{') - { - goto basic_json_parser_26; - } - goto basic_json_parser_4; } else { - if (yych <= '}') + if (yych <= '|') { - goto basic_json_parser_28; + if (yych == '{') + { + goto basic_json_parser_26; + } + goto basic_json_parser_4; } - if (yych == 0xEF) + else { - goto basic_json_parser_30; + if (yych <= '}') + { + goto basic_json_parser_28; + } + if (yych == 0xEF) + { + goto basic_json_parser_30; + } + goto basic_json_parser_4; } - goto basic_json_parser_4; } } - } basic_json_parser_2: - ++m_cursor; - { - return token_type::end_of_input; - } + ++m_cursor; + { + last_token_type = token_type::end_of_input; + break; + } basic_json_parser_4: - ++m_cursor; + ++m_cursor; basic_json_parser_5: - { - return token_type::parse_error; - } + { + last_token_type = token_type::parse_error; + break; + } basic_json_parser_6: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yybm[0 + yych] & 32) - { - goto basic_json_parser_6; - } - { - return scan(); - } + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yybm[0 + yych] & 32) + { + goto basic_json_parser_6; + } + { + continue; + } basic_json_parser_9: - yyaccept = 0; - yych = *(m_marker = ++m_cursor); - if (yych <= 0x0F) - { + yyaccept = 0; + yych = *(m_marker = ++m_cursor); + if (yych <= 0x1F) + { + goto basic_json_parser_5; + } + goto basic_json_parser_32; +basic_json_parser_10: + ++m_cursor; + { + last_token_type = token_type::value_separator; + break; + } +basic_json_parser_12: + yych = *++m_cursor; + if (yych <= '/') + { + goto basic_json_parser_5; + } + if (yych <= '0') + { + goto basic_json_parser_13; + } + if (yych <= '9') + { + goto basic_json_parser_15; + } goto basic_json_parser_5; - } - goto basic_json_parser_32; -basic_json_parser_10: - ++m_cursor; - { - return token_type::value_separator; - } -basic_json_parser_12: - yych = *++m_cursor; - if (yych <= '/') - { - goto basic_json_parser_5; - } - if (yych <= '0') - { - goto basic_json_parser_13; - } - if (yych <= '9') - { - goto basic_json_parser_15; - } - goto basic_json_parser_5; basic_json_parser_13: - yyaccept = 1; - yych = *(m_marker = ++m_cursor); - if (yych <= 'D') - { - if (yych == '.') - { - goto basic_json_parser_37; - } - } - else - { - if (yych <= 'E') - { - goto basic_json_parser_38; - } - if (yych == 'e') - { - goto basic_json_parser_38; - } - } + yyaccept = 1; + yych = *(m_marker = ++m_cursor); + if (yych <= 'D') + { + if (yych == '.') + { + goto basic_json_parser_37; + } + } + else + { + if (yych <= 'E') + { + goto basic_json_parser_38; + } + if (yych == 'e') + { + goto basic_json_parser_38; + } + } basic_json_parser_14: - { - return token_type::value_number; - } + { + last_token_type = token_type::value_number; + break; + } basic_json_parser_15: - yyaccept = 1; - m_marker = ++m_cursor; - if ((m_limit - m_cursor) < 3) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yybm[0 + yych] & 64) - { - goto basic_json_parser_15; - } - if (yych <= 'D') - { - if (yych == '.') - { - goto basic_json_parser_37; - } - goto basic_json_parser_14; - } - else - { - if (yych <= 'E') - { - goto basic_json_parser_38; - } - if (yych == 'e') - { - goto basic_json_parser_38; - } - goto basic_json_parser_14; - } -basic_json_parser_17: - ++m_cursor; - { - return token_type::name_separator; - } -basic_json_parser_19: - ++m_cursor; - { - return token_type::begin_array; - } -basic_json_parser_21: - ++m_cursor; - { - return token_type::end_array; - } -basic_json_parser_23: - yyaccept = 0; - yych = *(m_marker = ++m_cursor); - if (yych == 'a') - { - goto basic_json_parser_39; - } - goto basic_json_parser_5; -basic_json_parser_24: - yyaccept = 0; - yych = *(m_marker = ++m_cursor); - if (yych == 'u') - { - goto basic_json_parser_40; - } - goto basic_json_parser_5; -basic_json_parser_25: - yyaccept = 0; - yych = *(m_marker = ++m_cursor); - if (yych == 'r') - { - goto basic_json_parser_41; - } - goto basic_json_parser_5; -basic_json_parser_26: - ++m_cursor; - { - return token_type::begin_object; - } -basic_json_parser_28: - ++m_cursor; - { - return token_type::end_object; - } -basic_json_parser_30: - yyaccept = 0; - yych = *(m_marker = ++m_cursor); - if (yych == 0xBB) - { - goto basic_json_parser_42; - } - goto basic_json_parser_5; -basic_json_parser_31: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; -basic_json_parser_32: - if (yybm[0 + yych] & 128) - { - goto basic_json_parser_31; - } - if (yych <= 0x0F) - { - goto basic_json_parser_33; - } - if (yych <= '"') - { - goto basic_json_parser_34; - } - goto basic_json_parser_36; -basic_json_parser_33: - m_cursor = m_marker; - if (yyaccept == 0) - { - goto basic_json_parser_5; - } - else - { - goto basic_json_parser_14; - } -basic_json_parser_34: - ++m_cursor; - { - return token_type::value_string; - } -basic_json_parser_36: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= 'e') - { - if (yych <= '/') - { - if (yych == '"') + yyaccept = 1; + m_marker = ++m_cursor; + if ((m_limit - m_cursor) < 3) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yybm[0 + yych] & 64) + { + goto basic_json_parser_15; + } + if (yych <= 'D') + { + if (yych == '.') { - goto basic_json_parser_31; + goto basic_json_parser_37; } - if (yych <= '.') - { - goto basic_json_parser_33; - } - goto basic_json_parser_31; + goto basic_json_parser_14; } else { - if (yych <= '\\') + if (yych <= 'E') + { + goto basic_json_parser_38; + } + if (yych == 'e') { - if (yych <= '[') + goto basic_json_parser_38; + } + goto basic_json_parser_14; + } +basic_json_parser_17: + ++m_cursor; + { + last_token_type = token_type::name_separator; + break; + } +basic_json_parser_19: + ++m_cursor; + { + last_token_type = token_type::begin_array; + break; + } +basic_json_parser_21: + ++m_cursor; + { + last_token_type = token_type::end_array; + break; + } +basic_json_parser_23: + yyaccept = 0; + yych = *(m_marker = ++m_cursor); + if (yych == 'a') + { + goto basic_json_parser_39; + } + goto basic_json_parser_5; +basic_json_parser_24: + yyaccept = 0; + yych = *(m_marker = ++m_cursor); + if (yych == 'u') + { + goto basic_json_parser_40; + } + goto basic_json_parser_5; +basic_json_parser_25: + yyaccept = 0; + yych = *(m_marker = ++m_cursor); + if (yych == 'r') + { + goto basic_json_parser_41; + } + goto basic_json_parser_5; +basic_json_parser_26: + ++m_cursor; + { + last_token_type = token_type::begin_object; + break; + } +basic_json_parser_28: + ++m_cursor; + { + last_token_type = token_type::end_object; + break; + } +basic_json_parser_30: + yyaccept = 0; + yych = *(m_marker = ++m_cursor); + if (yych == 0xBB) + { + goto basic_json_parser_42; + } + goto basic_json_parser_5; +basic_json_parser_31: + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; +basic_json_parser_32: + if (yybm[0 + yych] & 128) + { + goto basic_json_parser_31; + } + if (yych <= 0x1F) + { + goto basic_json_parser_33; + } + if (yych <= '"') + { + goto basic_json_parser_34; + } + goto basic_json_parser_36; +basic_json_parser_33: + m_cursor = m_marker; + if (yyaccept == 0) + { + goto basic_json_parser_5; + } + else + { + goto basic_json_parser_14; + } +basic_json_parser_34: + ++m_cursor; + { + last_token_type = token_type::value_string; + break; + } +basic_json_parser_36: + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yych <= 'e') + { + if (yych <= '/') + { + if (yych == '"') + { + goto basic_json_parser_31; + } + if (yych <= '.') { goto basic_json_parser_33; } @@ -7739,33 +8177,33 @@ } else { - if (yych == 'b') + if (yych <= '\\') + { + if (yych <= '[') + { + goto basic_json_parser_33; + } + goto basic_json_parser_31; + } + else + { + if (yych == 'b') + { + goto basic_json_parser_31; + } + goto basic_json_parser_33; + } + } + } + else + { + if (yych <= 'q') + { + if (yych <= 'f') { goto basic_json_parser_31; } - goto basic_json_parser_33; - } - } - } - else - { - if (yych <= 'q') - { - if (yych <= 'f') - { - goto basic_json_parser_31; - } - if (yych == 'n') - { - goto basic_json_parser_31; - } - goto basic_json_parser_33; - } - else - { - if (yych <= 's') - { - if (yych <= 'r') + if (yych == 'n') { goto basic_json_parser_31; } @@ -7773,361 +8211,423 @@ } else { - if (yych <= 't') + if (yych <= 's') { - goto basic_json_parser_31; + if (yych <= 'r') + { + goto basic_json_parser_31; + } + goto basic_json_parser_33; } - if (yych <= 'u') + else { - goto basic_json_parser_43; + if (yych <= 't') + { + goto basic_json_parser_31; + } + if (yych <= 'u') + { + goto basic_json_parser_43; + } + goto basic_json_parser_33; } - goto basic_json_parser_33; } } - } basic_json_parser_37: - yych = *++m_cursor; - if (yych <= '/') - { - goto basic_json_parser_33; - } - if (yych <= '9') - { - goto basic_json_parser_44; - } - goto basic_json_parser_33; -basic_json_parser_38: - yych = *++m_cursor; - if (yych <= ',') - { - if (yych == '+') - { - goto basic_json_parser_46; - } - goto basic_json_parser_33; - } - else - { - if (yych <= '-') - { - goto basic_json_parser_46; - } + yych = *++m_cursor; if (yych <= '/') { goto basic_json_parser_33; } if (yych <= '9') { - goto basic_json_parser_47; + goto basic_json_parser_44; + } + goto basic_json_parser_33; +basic_json_parser_38: + yych = *++m_cursor; + if (yych <= ',') + { + if (yych == '+') + { + goto basic_json_parser_46; + } + goto basic_json_parser_33; + } + else + { + if (yych <= '-') + { + goto basic_json_parser_46; + } + if (yych <= '/') + { + goto basic_json_parser_33; + } + if (yych <= '9') + { + goto basic_json_parser_47; + } + goto basic_json_parser_33; + } +basic_json_parser_39: + yych = *++m_cursor; + if (yych == 'l') + { + goto basic_json_parser_49; + } + goto basic_json_parser_33; +basic_json_parser_40: + yych = *++m_cursor; + if (yych == 'l') + { + goto basic_json_parser_50; + } + goto basic_json_parser_33; +basic_json_parser_41: + yych = *++m_cursor; + if (yych == 'u') + { + goto basic_json_parser_51; + } + goto basic_json_parser_33; +basic_json_parser_42: + yych = *++m_cursor; + if (yych == 0xBF) + { + goto basic_json_parser_52; } goto basic_json_parser_33; - } -basic_json_parser_39: - yych = *++m_cursor; - if (yych == 'l') - { - goto basic_json_parser_49; - } - goto basic_json_parser_33; -basic_json_parser_40: - yych = *++m_cursor; - if (yych == 'l') - { - goto basic_json_parser_50; - } - goto basic_json_parser_33; -basic_json_parser_41: - yych = *++m_cursor; - if (yych == 'u') - { - goto basic_json_parser_51; - } - goto basic_json_parser_33; -basic_json_parser_42: - yych = *++m_cursor; - if (yych == 0xBF) - { - goto basic_json_parser_52; - } - goto basic_json_parser_33; basic_json_parser_43: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= '@') - { + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yych <= '@') + { + if (yych <= '/') + { + goto basic_json_parser_33; + } + if (yych <= '9') + { + goto basic_json_parser_54; + } + goto basic_json_parser_33; + } + else + { + if (yych <= 'F') + { + goto basic_json_parser_54; + } + if (yych <= '`') + { + goto basic_json_parser_33; + } + if (yych <= 'f') + { + goto basic_json_parser_54; + } + goto basic_json_parser_33; + } +basic_json_parser_44: + yyaccept = 1; + m_marker = ++m_cursor; + if ((m_limit - m_cursor) < 3) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yych <= 'D') + { + if (yych <= '/') + { + goto basic_json_parser_14; + } + if (yych <= '9') + { + goto basic_json_parser_44; + } + goto basic_json_parser_14; + } + else + { + if (yych <= 'E') + { + goto basic_json_parser_38; + } + if (yych == 'e') + { + goto basic_json_parser_38; + } + goto basic_json_parser_14; + } +basic_json_parser_46: + yych = *++m_cursor; if (yych <= '/') { goto basic_json_parser_33; } - if (yych <= '9') - { - goto basic_json_parser_54; - } - goto basic_json_parser_33; - } - else - { - if (yych <= 'F') - { - goto basic_json_parser_54; - } - if (yych <= '`') + if (yych >= ':') { goto basic_json_parser_33; } - if (yych <= 'f') - { - goto basic_json_parser_54; - } - goto basic_json_parser_33; - } -basic_json_parser_44: - yyaccept = 1; - m_marker = ++m_cursor; - if ((m_limit - m_cursor) < 3) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= 'D') - { +basic_json_parser_47: + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; if (yych <= '/') { goto basic_json_parser_14; } if (yych <= '9') { - goto basic_json_parser_44; - } - goto basic_json_parser_14; - } - else - { - if (yych <= 'E') - { - goto basic_json_parser_38; - } - if (yych == 'e') - { - goto basic_json_parser_38; + goto basic_json_parser_47; } goto basic_json_parser_14; - } -basic_json_parser_46: - yych = *++m_cursor; - if (yych <= '/') - { - goto basic_json_parser_33; - } - if (yych >= ':') - { +basic_json_parser_49: + yych = *++m_cursor; + if (yych == 's') + { + goto basic_json_parser_55; + } goto basic_json_parser_33; - } -basic_json_parser_47: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= '/') - { - goto basic_json_parser_14; - } - if (yych <= '9') - { - goto basic_json_parser_47; - } - goto basic_json_parser_14; -basic_json_parser_49: - yych = *++m_cursor; - if (yych == 's') - { - goto basic_json_parser_55; - } - goto basic_json_parser_33; basic_json_parser_50: - yych = *++m_cursor; - if (yych == 'l') - { - goto basic_json_parser_56; - } - goto basic_json_parser_33; + yych = *++m_cursor; + if (yych == 'l') + { + goto basic_json_parser_56; + } + goto basic_json_parser_33; basic_json_parser_51: - yych = *++m_cursor; - if (yych == 'e') - { - goto basic_json_parser_58; - } - goto basic_json_parser_33; -basic_json_parser_52: - ++m_cursor; - { - return scan(); - } -basic_json_parser_54: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= '@') - { - if (yych <= '/') - { - goto basic_json_parser_33; - } - if (yych <= '9') - { - goto basic_json_parser_60; + yych = *++m_cursor; + if (yych == 'e') + { + goto basic_json_parser_58; } goto basic_json_parser_33; - } - else - { - if (yych <= 'F') - { - goto basic_json_parser_60; - } - if (yych <= '`') - { +basic_json_parser_52: + ++m_cursor; + { + continue; + } +basic_json_parser_54: + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yych <= '@') + { + if (yych <= '/') + { + goto basic_json_parser_33; + } + if (yych <= '9') + { + goto basic_json_parser_60; + } goto basic_json_parser_33; } - if (yych <= 'f') - { - goto basic_json_parser_60; + else + { + if (yych <= 'F') + { + goto basic_json_parser_60; + } + if (yych <= '`') + { + goto basic_json_parser_33; + } + if (yych <= 'f') + { + goto basic_json_parser_60; + } + goto basic_json_parser_33; + } +basic_json_parser_55: + yych = *++m_cursor; + if (yych == 'e') + { + goto basic_json_parser_61; } goto basic_json_parser_33; - } -basic_json_parser_55: - yych = *++m_cursor; - if (yych == 'e') - { - goto basic_json_parser_61; - } - goto basic_json_parser_33; basic_json_parser_56: - ++m_cursor; - { - return token_type::literal_null; - } + ++m_cursor; + { + last_token_type = token_type::literal_null; + break; + } basic_json_parser_58: - ++m_cursor; - { - return token_type::literal_true; - } + ++m_cursor; + { + last_token_type = token_type::literal_true; + break; + } basic_json_parser_60: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= '@') - { - if (yych <= '/') - { + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yych <= '@') + { + if (yych <= '/') + { + goto basic_json_parser_33; + } + if (yych <= '9') + { + goto basic_json_parser_63; + } goto basic_json_parser_33; } - if (yych <= '9') - { - goto basic_json_parser_63; - } - goto basic_json_parser_33; - } - else - { - if (yych <= 'F') - { - goto basic_json_parser_63; - } - if (yych <= '`') - { + else + { + if (yych <= 'F') + { + goto basic_json_parser_63; + } + if (yych <= '`') + { + goto basic_json_parser_33; + } + if (yych <= 'f') + { + goto basic_json_parser_63; + } goto basic_json_parser_33; } - if (yych <= 'f') - { - goto basic_json_parser_63; - } - goto basic_json_parser_33; - } basic_json_parser_61: - ++m_cursor; - { - return token_type::literal_false; - } + ++m_cursor; + { + last_token_type = token_type::literal_false; + break; + } basic_json_parser_63: - ++m_cursor; - if (m_limit <= m_cursor) - { - yyfill(); // LCOV_EXCL_LINE; - } - yych = *m_cursor; - if (yych <= '@') - { - if (yych <= '/') - { + ++m_cursor; + if (m_limit <= m_cursor) + { + fill_line_buffer(); + } + yych = *m_cursor; + if (yych <= '@') + { + if (yych <= '/') + { + goto basic_json_parser_33; + } + if (yych <= '9') + { + goto basic_json_parser_31; + } + goto basic_json_parser_33; + } + else + { + if (yych <= 'F') + { + goto basic_json_parser_31; + } + if (yych <= '`') + { + goto basic_json_parser_33; + } + if (yych <= 'f') + { + goto basic_json_parser_31; + } goto basic_json_parser_33; } - if (yych <= '9') - { - goto basic_json_parser_31; - } - goto basic_json_parser_33; - } - else - { - if (yych <= 'F') - { - goto basic_json_parser_31; - } - if (yych <= '`') - { - goto basic_json_parser_33; - } - if (yych <= 'f') - { - goto basic_json_parser_31; - } - goto basic_json_parser_33; - } - } - - } - - /// append data from the stream to the internal buffer - void yyfill() noexcept - { - if (m_stream == nullptr or not * m_stream) - { - return; - } - + } + + } + + return last_token_type; + } + + /*! + @brief append data from the stream to the line buffer + + This function is called by the scan() function when the end of the + buffer (`m_limit`) is reached and the `m_cursor` pointer cannot be + incremented without leaving the limits of the line buffer. Note re2c + decides when to call this function. + + If the lexer reads from contiguous storage, there is no trailing null + byte. Therefore, this function must make sure to add these padding + null bytes. + + If the lexer reads from an input stream, this function reads the next + line of the input. + + @pre + p p p p p p u u u u u x . . . . . . + ^ ^ ^ ^ + m_content m_start | m_limit + m_cursor + + @post + u u u u u x x x x x x x . . . . . . + ^ ^ ^ + | m_cursor m_limit + m_start + m_content + */ + void fill_line_buffer() + { + // number of processed characters (p) const auto offset_start = m_start - m_content; - const auto offset_marker = m_marker - m_start; + // offset for m_marker wrt. to m_start + const auto offset_marker = (m_marker == nullptr) ? 0 : m_marker - m_start; + // number of unprocessed characters (u) const auto offset_cursor = m_cursor - m_start; - m_buffer.erase(0, static_cast(offset_start)); - std::string line; - assert(m_stream != nullptr); - std::getline(*m_stream, line); - m_buffer += "\n" + line; // add line with newline symbol - - m_content = reinterpret_cast(m_buffer.c_str()); + // no stream is used or end of file is reached + if (m_stream == nullptr or m_stream->eof()) + { + // copy unprocessed characters to line buffer + m_line_buffer.clear(); + for (m_cursor = m_start; m_cursor != m_limit; ++m_cursor) + { + m_line_buffer.append(1, static_cast(*m_cursor)); + } + + // append 5 characters (size of longest keyword "false") to + // make sure that there is sufficient space between m_cursor + // and m_limit + m_line_buffer.append(5, '\0'); + } + else + { + // delete processed characters from line buffer + m_line_buffer.erase(0, static_cast(offset_start)); + // read next line from input stream + std::string line; + std::getline(*m_stream, line); + // add line with newline symbol to the line buffer + m_line_buffer += line + "\n"; + } + + // set pointers + m_content = reinterpret_cast(m_line_buffer.c_str()); assert(m_content != nullptr); m_start = m_content; m_marker = m_start + offset_marker; m_cursor = m_start + offset_cursor; - m_limit = m_start + m_buffer.size() - 1; + m_limit = m_start + m_line_buffer.size(); } /// return string representation of last read token - string_t get_token() const + string_t get_token_string() const { assert(m_start != nullptr); return string_t(reinterpret_cast(m_start), @@ -8152,12 +8652,49 @@ of the construction of the values. 2. Unescaped characters are copied as is. + @pre `m_cursor - m_start >= 2`, meaning the length of the last token + is at least 2 bytes which is trivially true for any string (which + consists of at least two quotes). + + " c1 c2 c3 ... " + ^ ^ + m_start m_cursor + + @complexity Linear in the length of the string.\n + + Lemma: The loop body will always terminate.\n + + Proof (by contradiction): Assume the loop body does not terminate. As + the loop body does not contain another loop, one of the called + functions must never return. The called functions are `std::strtoul` + and to_unicode. Neither function can loop forever, so the loop body + will never loop forever which contradicts the assumption that the loop + body does not terminate, q.e.d.\n + + Lemma: The loop condition for the for loop is eventually false.\n + + Proof (by contradiction): Assume the loop does not terminate. Due to + the above lemma, this can only be due to a tautological loop + condition; that is, the loop condition i < m_cursor - 1 must always be + true. Let x be the change of i for any loop iteration. Then + m_start + 1 + x < m_cursor - 1 must hold to loop indefinitely. This + can be rephrased to m_cursor - m_start - 2 > x. With the + precondition, we x <= 0, meaning that the loop condition holds + indefinitly if i is always decreased. However, observe that the value + of i is strictly increasing with each iteration, as it is incremented + by 1 in the iteration expression and never decremented inside the loop + body. Hence, the loop condition will eventually be false which + contradicts the assumption that the loop condition is a tautology, + q.e.d. + @return string value of current token without opening and closing quotes @throw std::out_of_range if to_unicode fails */ string_t get_string() const { + assert(m_cursor - m_start >= 2); + string_t result; result.reserve(static_cast(m_cursor - m_start - 2)); @@ -8273,11 +8810,6 @@ the number @return the floating point number - - @bug This function uses `std::strtof`, `std::strtod`, or `std::strtold` - which use the current C locale to determine which character is used as - decimal point character. This may yield to parse errors if the locale - does not used `.`. */ long double str_to_float_t(long double* /* type */, char** endptr) const { @@ -8395,7 +8927,7 @@ if (type != value_t::number_float) { // multiply last value by ten and add the new digit - auto temp = value * 10 + *curptr - 0x30; + auto temp = value * 10 + *curptr - '0'; // test for overflow if (temp < value || temp > max) @@ -8433,8 +8965,8 @@ private: /// optional input stream std::istream* m_stream = nullptr; - /// the buffer - string_t m_buffer; + /// line buffer buffer for m_stream + string_t m_line_buffer {}; /// the buffer pointer const lexer_char_t* m_content = nullptr; /// pointer to the beginning of the current symbol @@ -8445,6 +8977,8 @@ const lexer_char_t* m_cursor = nullptr; /// pointer to the end of the buffer const lexer_char_t* m_limit = nullptr; + /// the last token type + token_type last_token_type = token_type::end_of_input; }; /*! @@ -8455,32 +8989,42 @@ class parser { public: - /// constructor for strings - parser(const string_t& s, parser_callback_t cb = nullptr) noexcept - : callback(cb), m_lexer(s) + /// a parser reading from a string literal + parser(const char* buff, const parser_callback_t cb = nullptr) + : callback(cb), + m_lexer(reinterpret_cast(buff), strlen(buff)) + {} + + /// a parser reading from an input stream + parser(std::istream& is, const parser_callback_t cb = nullptr) + : callback(cb), m_lexer(is) + {} + + /// a parser reading from an iterator range with contiguous storage + template::iterator_category, std::random_access_iterator_tag>::value + , int>::type + = 0> + parser(IteratorType first, IteratorType last, const parser_callback_t cb = nullptr) + : callback(cb), + m_lexer(reinterpret_cast(&(*first)), + static_cast(std::distance(first, last))) + {} + + /// public parser interface + basic_json parse() { // read first token get_token(); - } - - /// a parser reading from an input stream - parser(std::istream& _is, parser_callback_t cb = nullptr) noexcept - : callback(cb), m_lexer(&_is) - { - // read first token - get_token(); - } - - /// public parser interface - basic_json parse() - { + basic_json result = parse_internal(true); + result.assert_invariant(); expect(lexer::token_type::end_of_input); // return parser result and replace it with null in case the // top-level value was discarded by the callback function - return result.is_discarded() ? basic_json() : result; + return result.is_discarded() ? basic_json() : std::move(result); } private: @@ -8493,11 +9037,12 @@ { case lexer::token_type::begin_object: { - if (keep and (not callback or (keep = callback(depth++, parse_event_t::object_start, result)))) + if (keep and (not callback + or ((keep = callback(depth++, parse_event_t::object_start, result)) != 0))) { // explicitly set result to object to cope with {} result.m_type = value_t::object; - result.m_value = json_value(value_t::object); + result.m_value = value_t::object; } // read next token @@ -8571,11 +9116,12 @@ case lexer::token_type::begin_array: { - if (keep and (not callback or (keep = callback(depth++, parse_event_t::array_start, result)))) + if (keep and (not callback + or ((keep = callback(depth++, parse_event_t::array_start, result)) != 0))) { // explicitly set result to object to cope with [] result.m_type = value_t::array; - result.m_value = json_value(value_t::array); + result.m_value = value_t::array; } // read next token @@ -8677,7 +9223,7 @@ } /// get next token from lexer - typename lexer::token_type get_token() noexcept + typename lexer::token_type get_token() { last_token = m_lexer.scan(); return last_token; @@ -8688,7 +9234,8 @@ if (t != last_token) { std::string error_msg = "parse error - unexpected "; - error_msg += (last_token == lexer::token_type::parse_error ? ("'" + m_lexer.get_token() + "'") : + error_msg += (last_token == lexer::token_type::parse_error ? ("'" + m_lexer.get_token_string() + + "'") : lexer::token_type_name(last_token)); error_msg += "; expected " + lexer::token_type_name(t); throw std::invalid_argument(error_msg); @@ -8700,7 +9247,8 @@ if (t == last_token) { std::string error_msg = "parse error - unexpected "; - error_msg += (last_token == lexer::token_type::parse_error ? ("'" + m_lexer.get_token() + "'") : + error_msg += (last_token == lexer::token_type::parse_error ? ("'" + m_lexer.get_token_string() + + "'") : lexer::token_type_name(last_token)); throw std::invalid_argument(error_msg); } @@ -8710,7 +9258,7 @@ /// current level of recursion int depth = 0; /// callback function - parser_callback_t callback; + const parser_callback_t callback = nullptr; /// the type of the last read token typename lexer::token_type last_token = lexer::token_type::uninitialized; /// the lexer @@ -8778,14 +9326,12 @@ */ std::string to_string() const noexcept { - std::string result; - - for (const auto& reference_token : reference_tokens) - { - result += "/" + escape(reference_token); - } - - return result; + return std::accumulate(reference_tokens.begin(), + reference_tokens.end(), std::string{}, + [](const std::string & a, const std::string & b) + { + return a + "/" + escape(b); + }); } /// @copydoc to_string() @@ -8828,6 +9374,8 @@ /*! @brief create and return a reference to the pointed to value + + @complexity Linear in the number of reference tokens. */ reference get_and_create(reference j) const { @@ -9083,7 +9631,7 @@ } /// split the string input to reference tokens - static std::vector split(std::string reference_string) + static std::vector split(const std::string& reference_string) { std::vector result; @@ -9149,7 +9697,7 @@ @param[in,out] s the string to manipulate @param[in] f the substring to replace with @a t - @param[out] t the string to replace @a f + @param[in] t the string to replace @a f @return The string @a s where all occurrences of @a f are replaced with @a t. @@ -9568,7 +10116,7 @@ json_pointer top_pointer = ptr.top(); if (top_pointer != ptr) { - basic_json& x = result.at(top_pointer); + result.at(top_pointer); } // get reference to parent of JSON pointer ptr @@ -9811,7 +10359,7 @@ */ static basic_json diff(const basic_json& source, const basic_json& target, - std::string path = "") + const std::string& path = "") { // the patch basic_json result(value_t::array); @@ -9973,7 +10521,7 @@ @since version 1.0.0 */ -template <> +template<> inline void swap(nlohmann::json& j1, nlohmann::json& j2) noexcept( is_nothrow_move_constructible::value and @@ -9984,7 +10532,7 @@ } /// hash value for JSON objects -template <> +template<> struct hash { /*! @@ -10005,7 +10553,7 @@ @brief user-defined string literal for JSON values This operator implements a user-defined string literal for JSON objects. It -can be used by adding \p "_json" to a string literal and returns a JSON object +can be used by adding `"_json"` to a string literal and returns a JSON object if no parse error occurred. @param[in] s a string representation of a JSON object @@ -10015,12 +10563,19 @@ */ inline nlohmann::json operator "" _json(const char* s, std::size_t) { - return nlohmann::json::parse(reinterpret_cast(s)); + return nlohmann::json::parse(s); } /*! @brief user-defined string literal for JSON pointer +This operator implements a user-defined string literal for JSON Pointers. It +can be used by adding `"_json"` to a string literal and returns a JSON pointer +object if no parse error occurred. + +@param[in] s a string representation of a JSON Pointer +@return a JSON pointer object + @since version 2.0.0 */ inline nlohmann::json::json_pointer operator "" _json_pointer(const char* s, std::size_t)