Skip to content

J
json
Commit 2d8b3628 by Niels
Options

further cleanup

parent 2fc82358
Showing with 255 additions and 201 deletions
Doxyfile
...@@ -13,7 +13,7 @@ CREATE_SUBDIRS = NO ...@@ -13,7 +13,7 @@ CREATE_SUBDIRS = NO
ALLOW_UNICODE_NAMES = NO ALLOW_UNICODE_NAMES = NO
OUTPUT_LANGUAGE = English OUTPUT_LANGUAGE = English
BRIEF_MEMBER_DESC = YES BRIEF_MEMBER_DESC = YES
REPEAT_BRIEF = YES REPEAT_BRIEF = NO
ABBREVIATE_BRIEF = ABBREVIATE_BRIEF =
ALWAYS_DETAILED_SEC = YES ALWAYS_DETAILED_SEC = YES
INLINE_INHERITED_MEMB = NO INLINE_INHERITED_MEMB = NO
......
src/json.hpp
...@@ -24,14 +24,8 @@ ...@@ -24,14 +24,8 @@
#include <utility> #include <utility>
#include <vector> #include <vector>
/*!
- ObjectType trick from http://stackoverflow.com/a/9860911
*/
/* /*
template<typename C, typename=void> - ObjectType trick from http://stackoverflow.com/a/9860911
struct container_resizable : std::false_type {};
template<typename C>
struct container_resizable<C, decltype(C().resize(0))> : std::true_type {};
*/ */
/*! /*!
...@@ -55,10 +49,12 @@ namespace nlohmann ...@@ -55,10 +49,12 @@ namespace nlohmann
(@c int64_t by default) (@c int64_t by default)
@tparam NumberFloatType type for JSON floating-point numbers @tparam NumberFloatType type for JSON floating-point numbers
(@c double by default) (@c double by default)
@see RFC 7159 <http://rfc7159.net/rfc7159>
@see ECMA 404 <http://www.ecma-international.org/publications/standards/Ecma-404.htm>
*/ */
template < template <
template<typename U, typename V, typename... Args> class ObjectType = std::map, template<typename U, typename V, typename... Args> class ObjectType = std::map,
//template<typename... Args> class ArrayType = std::vector,
template<typename U, typename... Args> class ArrayType = std::vector, template<typename U, typename... Args> class ArrayType = std::vector,
class StringType = std::string, class StringType = std::string,
class BooleanType = bool, class BooleanType = bool,
...@@ -72,6 +68,7 @@ class basic_json ...@@ -72,6 +68,7 @@ class basic_json
// container types // // container types //
///////////////////// /////////////////////
// forward declarations
class iterator; class iterator;
class const_iterator; class const_iterator;
...@@ -98,6 +95,7 @@ class basic_json ...@@ -98,6 +95,7 @@ class basic_json
/// a const reverse iterator for a basic_json container /// a const reverse iterator for a basic_json container
using const_reverse_iterator = std::reverse_iterator<const_iterator>; using const_reverse_iterator = std::reverse_iterator<const_iterator>;
/////////////////////////// ///////////////////////////
// JSON value data types // // JSON value data types //
/////////////////////////// ///////////////////////////
...@@ -162,20 +160,13 @@ class basic_json ...@@ -162,20 +160,13 @@ class basic_json
/// JSON value type enumeration /// JSON value type enumeration
enum class value_t : uint8_t enum class value_t : uint8_t
{ {
/// null value null, ///< null value
null, object, ///< object (unordered set of name/value pairs)
/// object (unordered set of name/value pairs) array, ///< array (ordered collection of values)
object, string, ///< string value
/// array (ordered collection of values) boolean, ///< boolean value
array, number_integer, ///< number value (integer)
/// string value number_float ///< number value (floating-point)
string,
/// boolean value
boolean,
/// number value (integer)
number_integer,
/// number value (floating-point)
number_float
}; };
...@@ -524,7 +515,7 @@ class basic_json ...@@ -524,7 +515,7 @@ class basic_json
@see https://docs.python.org/2/library/json.html#json.dump @see https://docs.python.org/2/library/json.html#json.dump
*/ */
inline string_t dump(int indent = -1) const noexcept inline string_t dump(const int indent = -1) const noexcept
{ {
if (indent >= 0) if (indent >= 0)
{ {
...@@ -833,12 +824,24 @@ class basic_json ...@@ -833,12 +824,24 @@ class basic_json
return reverse_iterator(end()); return reverse_iterator(end());
} }
/// returns a reverse iterator to the beginning
inline const_reverse_iterator rbegin() const noexcept
{
return const_reverse_iterator(end());
}
/// returns a reverse iterator to the end /// returns a reverse iterator to the end
inline reverse_iterator rend() noexcept inline reverse_iterator rend() noexcept
{ {
return reverse_iterator(begin()); return reverse_iterator(begin());
} }
/// returns a reverse iterator to the end
inline const_reverse_iterator rend() const noexcept
{
return const_reverse_iterator(begin());
}
/// returns a reverse iterator to the beginning /// returns a reverse iterator to the beginning
inline const_reverse_iterator crbegin() const noexcept inline const_reverse_iterator crbegin() const noexcept
{ {
...@@ -1335,7 +1338,7 @@ class basic_json ...@@ -1335,7 +1338,7 @@ class basic_json
///////////////////// /////////////////////
/// deserialize from string /// deserialize from string
static basic_json parse(const std::string& s) static basic_json parse(const string_t& s)
{ {
return parser(s).parse(); return parser(s).parse();
} }
...@@ -1501,7 +1504,7 @@ class basic_json ...@@ -1501,7 +1504,7 @@ class basic_json
@param currentIndent the current indent level (only used internally) @param currentIndent the current indent level (only used internally)
*/ */
inline string_t dump(const bool prettyPrint, const unsigned int indentStep, inline string_t dump(const bool prettyPrint, const unsigned int indentStep,
unsigned int currentIndent = 0) const noexcept const unsigned int currentIndent = 0) const noexcept
{ {
// variable to hold indentation for recursive calls // variable to hold indentation for recursive calls
auto new_indent = currentIndent; auto new_indent = currentIndent;
...@@ -1647,45 +1650,26 @@ class basic_json ...@@ -1647,45 +1650,26 @@ class basic_json
}; };
/// an iterator value /// an iterator value
template<typename array_iterator_t, typename object_iterator_t>
union internal_iterator union internal_iterator
{ {
/// iterator for JSON objects /// iterator for JSON objects
typename object_t::iterator object_iterator; object_iterator_t object_iterator;
/// iterator for JSON arrays /// iterator for JSON arrays
typename array_t::iterator array_iterator; array_iterator_t array_iterator;
/// generic iteraotr for all other value types /// generic iteraotr for all other value types
generic_iterator_value generic_iterator; generic_iterator_value generic_iterator;
/// default constructor /// default constructor
internal_iterator() : generic_iterator(generic_iterator_value::uninitialized) {} internal_iterator() : generic_iterator(generic_iterator_value::uninitialized) {}
/// constructor for object iterators /// constructor for object iterators
internal_iterator(typename object_t::iterator v) : object_iterator(v) {} internal_iterator(object_iterator_t v) : object_iterator(v) {}
/// constructor for array iterators /// constructor for array iterators
internal_iterator(typename array_t::iterator v) : array_iterator(v) {} internal_iterator(array_iterator_t v) : array_iterator(v) {}
/// constructor for generic iterators /// constructor for generic iterators
internal_iterator(generic_iterator_value v) : generic_iterator(v) {} internal_iterator(generic_iterator_value v) : generic_iterator(v) {}
}; };
/// a const iterator value
union internal_const_iterator
{
/// iterator for JSON objects
typename object_t::const_iterator object_iterator;
/// iterator for JSON arrays
typename array_t::const_iterator array_iterator;
/// generic iteraotr for all other value types
generic_iterator_value generic_iterator;
/// default constructor
internal_const_iterator() : generic_iterator(generic_iterator_value::uninitialized) {}
/// constructor for object iterators
internal_const_iterator(typename object_t::iterator v) : object_iterator(v) {}
/// constructor for array iterators
internal_const_iterator(typename array_t::iterator v) : array_iterator(v) {}
/// constructor for generic iterators
internal_const_iterator(generic_iterator_value v) : generic_iterator(v) {}
};
public: public:
/// a bidirectional iterator for the basic_json class /// a bidirectional iterator for the basic_json class
class iterator : public std::iterator<std::bidirectional_iterator_tag, basic_json> class iterator : public std::iterator<std::bidirectional_iterator_tag, basic_json>
...@@ -2042,7 +2026,7 @@ class basic_json ...@@ -2042,7 +2026,7 @@ class basic_json
/// associated JSON instance /// associated JSON instance
pointer m_object = nullptr; pointer m_object = nullptr;
/// the actual iterator of the associated instance /// the actual iterator of the associated instance
internal_iterator m_it; internal_iterator<typename array_t::iterator, typename object_t::iterator> m_it;
}; };
/// a const bidirectional iterator for the basic_json class /// a const bidirectional iterator for the basic_json class
...@@ -2425,49 +2409,68 @@ class basic_json ...@@ -2425,49 +2409,68 @@ class basic_json
/// associated JSON instance /// associated JSON instance
pointer m_object = nullptr; pointer m_object = nullptr;
/// the actual iterator of the associated instance /// the actual iterator of the associated instance
internal_const_iterator m_it; internal_iterator<typename array_t::const_iterator, typename object_t::const_iterator> m_it;
}; };
private: private:
//////////// //////////////////////
// parser // // lexer and parser //
//////////// //////////////////////
/*!
@brief lexical analysis
This class organizes the lexical analysis during JSON deserialization. The
core of it is a scanner generated by re2c <http://re2c.org> that processes
a buffer and recognizes tokens according to RFC 7159 and ECMA-404.
*/
class lexer class lexer
{ {
public: public:
/// token types for the parser /// token types for the parser
enum class token_type enum class token_type
{ {
uninitialized, uninitialized, ///< indicating the scanner is uninitialized
literal_true, literal_true, ///< the "true" literal
literal_false, literal_false, ///< the "false" literal
literal_null, literal_null, ///< the "null" literal
value_string, value_string, ///< a string - use get_string() for actual value
value_number, value_number, ///< a number - use get_number() for actual value
begin_array, begin_array, ///< the character for array begin "["
begin_object, begin_object, ///< the character for object begin "{"
end_array, end_array, ///< the character for array end "]"
end_object, end_object, ///< the character for object end "}"
name_separator, name_separator, ///< the name separator ":"
value_separator, value_separator, ///< the value separator ","
parse_error, parse_error, ///< indicating a parse error
end_of_input end_of_input ///< indicating the end of the input buffer
}; };
/// the char type to use in the lexer /// the char type to use in the lexer
using lexer_char_t = typename string_t::value_type; using lexer_char_t = typename string_t::value_type;
inline lexer(const typename string_t::value_type* s) : m_content(s) /// constructor with a given buffer
inline lexer(const string_t& s) noexcept
: m_content(s.c_str())
{ {
m_start = m_cursor = m_content; m_start = m_cursor = m_content;
m_limit = m_content + strlen(m_content); m_limit = m_content + s.size();
} }
/// default constructor
inline lexer() = default; inline lexer() = default;
inline static string_t to_unicode(const long codepoint) /*!
@brief create a string from a Unicode code point
@param codepoint the code point (must be in [0x0, 0x10ffff]
@return string representation of the code point
@exception std::out_of_range if code point is >0x10ffff
@see <http://en.wikipedia.org/wiki/UTF-8#Sample_code>
*/
inline static string_t to_unicode(const size_t codepoint)
{ {
string_t result; string_t result;
...@@ -2505,7 +2508,8 @@ class basic_json ...@@ -2505,7 +2508,8 @@ class basic_json
return result; return result;
} }
inline static std::string token_type_name(token_type t) /// return name of values of type token_type
inline static std::string token_type_name(token_type t) noexcept
{ {
switch (t) switch (t)
{ {
...@@ -2544,16 +2548,16 @@ class basic_json ...@@ -2544,16 +2548,16 @@ class basic_json
This function implements a scanner for JSON. It is specified using This function implements a scanner for JSON. It is specified using
regular expressions that try to follow RFC 7159 and ECMA-404 as close regular expressions that try to follow RFC 7159 and ECMA-404 as close
as possible. These regular expressions are then translated into a as possible. These regular expressions are then translated into a
deterministic finite automaton (DFA) by the tool RE2C. As a result, the deterministic finite automaton (DFA) by the tool re2c
translated code for this function consists of a large block of code <http://re2c.org>. As a result, the translated code for this function
with goto jumps. consists of a large block of code with goto jumps.
@return the class of the next token read from the buffer @return the class of the next token read from the buffer
*/ */
inline token_type scan() inline token_type scan() noexcept
{ {
// pointer for backtracking information // pointer for backtracking information
const typename string_t::value_type* m_marker = nullptr; const lexer_char_t* m_marker = nullptr;
// remember the begin of the token // remember the begin of the token
m_start = m_cursor; m_start = m_cursor;
...@@ -3269,19 +3273,32 @@ basic_json_parser_59: ...@@ -3269,19 +3273,32 @@ basic_json_parser_59:
} }
inline string_t get_token() const /// return string representation of last read token
inline string_t get_token() const noexcept
{ {
return string_t(m_start, static_cast<size_t>(m_cursor - m_start)); return string_t(m_start, static_cast<size_t>(m_cursor - m_start));
} }
/*! /*!
The pointer m_start points to the opening quote of the string, and @brief return string value for string tokens
m_cursor past the closing quote of the string. We create a std::string
from the character after the opening quotes (m_start+1) until the The function iterates the characters between the opening and closing
character before the closing quotes (hence subtracting 2 characters quotes of the string value. The complete string is the range
from the pointer difference of the two pointers). [m_start,m_cursor). Consequently, we iterate from m_start+1 to
m_cursor-1.
We differentiate two cases:
1. Escaped characters. In this case, a new character is constructed
according to the nature of the escape. Some escapes create new
characters (e.g., @c "\\n" is replaced by @c "\n"), some are copied
as is (e.g., @c "\\\\"). Furthermore, Unicode escapes of the shape
@c "\\uxxxx" need special care. In this case, to_unicode takes care
of the construction of the values.
2. Unescaped characters are copied as is.
@return string value of current token without opening and closing quotes @return string value of current token without opening and closing quotes
@exception std::out_of_range if to_unicode fails
*/ */
inline string_t get_string() const inline string_t get_string() const
{ {
...@@ -3347,7 +3364,7 @@ basic_json_parser_59: ...@@ -3347,7 +3364,7 @@ basic_json_parser_59:
case 'u': case 'u':
{ {
// get code xxxx from \uxxxx // get code xxxx from \uxxxx
auto codepoint = strtol(i + 1, nullptr, 16); auto codepoint = std::strtoul(i + 1, nullptr, 16);
// add unicode character(s) // add unicode character(s)
result += to_unicode(codepoint); result += to_unicode(codepoint);
// skip the next four characters (\uxxxx) // skip the next four characters (\uxxxx)
...@@ -3367,19 +3384,29 @@ basic_json_parser_59: ...@@ -3367,19 +3384,29 @@ basic_json_parser_59:
return result; return result;
} }
/*!
@brief return number value for number tokens
This function translates the last token into a floating point number.
The pointer m_begin points to the beginning of the parsed number. We
pass this pointer to std::strtod which sets endptr to the first
character past the converted number. If this pointer is not the same as
m_cursor, then either more or less characters have been used during the
comparison. This can happen for inputs like "01" which will be treated
like number 0 followed by number 1.
@return the result of the number conversion or NAN if the conversion
read past the current token. The latter case needs to be treated by the
caller function.
@exception std::range_error if passed value is out of range
*/
inline number_float_t get_number() const inline number_float_t get_number() const
{ {
// The pointer m_begin points to the beginning of the parsed
// number. We pass this pointer to std::strtod which sets endptr to
// the first character past the converted number. If this pointer is
// not the same as m_cursor, then either more or less characters
// have been used during the comparison. This can happen for inputs
// like "01" which will be treated like number 0 followed by number
// 1.
// conversion // conversion
char* endptr; lexer_char_t* endptr;
const auto float_val = std::strtod(reinterpret_cast<const char*>(m_start), &endptr); const auto float_val = std::strtod(reinterpret_cast<const lexer_char_t*>(m_start),
&endptr);
// return float_val if the whole number was translated and NAN // return float_val if the whole number was translated and NAN
// otherwise // otherwise
...@@ -3388,20 +3415,20 @@ basic_json_parser_59: ...@@ -3388,20 +3415,20 @@ basic_json_parser_59:
private: private:
/// the buffer /// the buffer
const typename string_t::value_type* m_content = nullptr; const lexer_char_t* m_content = nullptr;
/// pointer to he beginning of the current symbol /// pointer to he beginning of the current symbol
const typename string_t::value_type* m_start = nullptr; const lexer_char_t* m_start = nullptr;
/// pointer to the current symbol /// pointer to the current symbol
const typename string_t::value_type* m_cursor = nullptr; const lexer_char_t* m_cursor = nullptr;
/// pointer to the end of the buffer /// pointer to the end of the buffer
const typename string_t::value_type* m_limit = nullptr; const lexer_char_t* m_limit = nullptr;
}; };
class parser class parser
{ {
public: public:
/// constructor for strings /// constructor for strings
inline parser(const string_t& s) : m_buffer(s), m_lexer(m_buffer.c_str()) inline parser(const string_t& s) : m_buffer(s), m_lexer(m_buffer)
{ {
// read first token // read first token
get_token(); get_token();
...@@ -3418,7 +3445,7 @@ basic_json_parser_59: ...@@ -3418,7 +3445,7 @@ basic_json_parser_59:
} }
// initializer lexer // initializer lexer
m_lexer = lexer(m_buffer.c_str()); m_lexer = lexer(m_buffer);
// read first token // read first token
get_token(); get_token();
......
src/json.hpp.re2c
...@@ -24,14 +24,8 @@ ...@@ -24,14 +24,8 @@
#include <utility> #include <utility>
#include <vector> #include <vector>
/*!
- ObjectType trick from http://stackoverflow.com/a/9860911
*/
/* /*
template<typename C, typename=void> - ObjectType trick from http://stackoverflow.com/a/9860911
struct container_resizable : std::false_type {};
template<typename C>
struct container_resizable<C, decltype(C().resize(0))> : std::true_type {};
*/ */
/*! /*!
...@@ -55,10 +49,12 @@ namespace nlohmann ...@@ -55,10 +49,12 @@ namespace nlohmann
(@c int64_t by default) (@c int64_t by default)
@tparam NumberFloatType type for JSON floating-point numbers @tparam NumberFloatType type for JSON floating-point numbers
(@c double by default) (@c double by default)
@see RFC 7159 <http://rfc7159.net/rfc7159>
@see ECMA 404 <http://www.ecma-international.org/publications/standards/Ecma-404.htm>
*/ */
template < template <
template<typename U, typename V, typename... Args> class ObjectType = std::map, template<typename U, typename V, typename... Args> class ObjectType = std::map,
//template<typename... Args> class ArrayType = std::vector,
template<typename U, typename... Args> class ArrayType = std::vector, template<typename U, typename... Args> class ArrayType = std::vector,
class StringType = std::string, class StringType = std::string,
class BooleanType = bool, class BooleanType = bool,
...@@ -72,6 +68,7 @@ class basic_json ...@@ -72,6 +68,7 @@ class basic_json
// container types // // container types //
///////////////////// /////////////////////
// forward declarations
class iterator; class iterator;
class const_iterator; class const_iterator;
...@@ -98,6 +95,7 @@ class basic_json ...@@ -98,6 +95,7 @@ class basic_json
/// a const reverse iterator for a basic_json container /// a const reverse iterator for a basic_json container
using const_reverse_iterator = std::reverse_iterator<const_iterator>; using const_reverse_iterator = std::reverse_iterator<const_iterator>;
/////////////////////////// ///////////////////////////
// JSON value data types // // JSON value data types //
/////////////////////////// ///////////////////////////
...@@ -162,20 +160,13 @@ class basic_json ...@@ -162,20 +160,13 @@ class basic_json
/// JSON value type enumeration /// JSON value type enumeration
enum class value_t : uint8_t enum class value_t : uint8_t
{ {
/// null value null, ///< null value
null, object, ///< object (unordered set of name/value pairs)
/// object (unordered set of name/value pairs) array, ///< array (ordered collection of values)
object, string, ///< string value
/// array (ordered collection of values) boolean, ///< boolean value
array, number_integer, ///< number value (integer)
/// string value number_float ///< number value (floating-point)
string,
/// boolean value
boolean,
/// number value (integer)
number_integer,
/// number value (floating-point)
number_float
}; };
...@@ -524,7 +515,7 @@ class basic_json ...@@ -524,7 +515,7 @@ class basic_json
@see https://docs.python.org/2/library/json.html#json.dump @see https://docs.python.org/2/library/json.html#json.dump
*/ */
inline string_t dump(int indent = -1) const noexcept inline string_t dump(const int indent = -1) const noexcept
{ {
if (indent >= 0) if (indent >= 0)
{ {
...@@ -833,12 +824,24 @@ class basic_json ...@@ -833,12 +824,24 @@ class basic_json
return reverse_iterator(end()); return reverse_iterator(end());
} }
/// returns a reverse iterator to the beginning
inline const_reverse_iterator rbegin() const noexcept
{
return const_reverse_iterator(end());
}
/// returns a reverse iterator to the end /// returns a reverse iterator to the end
inline reverse_iterator rend() noexcept inline reverse_iterator rend() noexcept
{ {
return reverse_iterator(begin()); return reverse_iterator(begin());
} }
/// returns a reverse iterator to the end
inline const_reverse_iterator rend() const noexcept
{
return const_reverse_iterator(begin());
}
/// returns a reverse iterator to the beginning /// returns a reverse iterator to the beginning
inline const_reverse_iterator crbegin() const noexcept inline const_reverse_iterator crbegin() const noexcept
{ {
...@@ -1335,7 +1338,7 @@ class basic_json ...@@ -1335,7 +1338,7 @@ class basic_json
///////////////////// /////////////////////
/// deserialize from string /// deserialize from string
static basic_json parse(const std::string& s) static basic_json parse(const string_t& s)
{ {
return parser(s).parse(); return parser(s).parse();
} }
...@@ -1501,7 +1504,7 @@ class basic_json ...@@ -1501,7 +1504,7 @@ class basic_json
@param currentIndent the current indent level (only used internally) @param currentIndent the current indent level (only used internally)
*/ */
inline string_t dump(const bool prettyPrint, const unsigned int indentStep, inline string_t dump(const bool prettyPrint, const unsigned int indentStep,
unsigned int currentIndent = 0) const noexcept const unsigned int currentIndent = 0) const noexcept
{ {
// variable to hold indentation for recursive calls // variable to hold indentation for recursive calls
auto new_indent = currentIndent; auto new_indent = currentIndent;
...@@ -1647,45 +1650,26 @@ class basic_json ...@@ -1647,45 +1650,26 @@ class basic_json
}; };
/// an iterator value /// an iterator value
template<typename array_iterator_t, typename object_iterator_t>
union internal_iterator union internal_iterator
{ {
/// iterator for JSON objects /// iterator for JSON objects
typename object_t::iterator object_iterator; object_iterator_t object_iterator;
/// iterator for JSON arrays /// iterator for JSON arrays
typename array_t::iterator array_iterator; array_iterator_t array_iterator;
/// generic iteraotr for all other value types /// generic iteraotr for all other value types
generic_iterator_value generic_iterator; generic_iterator_value generic_iterator;
/// default constructor /// default constructor
internal_iterator() : generic_iterator(generic_iterator_value::uninitialized) {} internal_iterator() : generic_iterator(generic_iterator_value::uninitialized) {}
/// constructor for object iterators /// constructor for object iterators
internal_iterator(typename object_t::iterator v) : object_iterator(v) {} internal_iterator(object_iterator_t v) : object_iterator(v) {}
/// constructor for array iterators /// constructor for array iterators
internal_iterator(typename array_t::iterator v) : array_iterator(v) {} internal_iterator(array_iterator_t v) : array_iterator(v) {}
/// constructor for generic iterators /// constructor for generic iterators
internal_iterator(generic_iterator_value v) : generic_iterator(v) {} internal_iterator(generic_iterator_value v) : generic_iterator(v) {}
}; };
/// a const iterator value
union internal_const_iterator
{
/// iterator for JSON objects
typename object_t::const_iterator object_iterator;
/// iterator for JSON arrays
typename array_t::const_iterator array_iterator;
/// generic iteraotr for all other value types
generic_iterator_value generic_iterator;
/// default constructor
internal_const_iterator() : generic_iterator(generic_iterator_value::uninitialized) {}
/// constructor for object iterators
internal_const_iterator(typename object_t::iterator v) : object_iterator(v) {}
/// constructor for array iterators
internal_const_iterator(typename array_t::iterator v) : array_iterator(v) {}
/// constructor for generic iterators
internal_const_iterator(generic_iterator_value v) : generic_iterator(v) {}
};
public: public:
/// a bidirectional iterator for the basic_json class /// a bidirectional iterator for the basic_json class
class iterator : public std::iterator<std::bidirectional_iterator_tag, basic_json> class iterator : public std::iterator<std::bidirectional_iterator_tag, basic_json>
...@@ -2042,7 +2026,7 @@ class basic_json ...@@ -2042,7 +2026,7 @@ class basic_json
/// associated JSON instance /// associated JSON instance
pointer m_object = nullptr; pointer m_object = nullptr;
/// the actual iterator of the associated instance /// the actual iterator of the associated instance
internal_iterator m_it; internal_iterator<typename array_t::iterator, typename object_t::iterator> m_it;
}; };
/// a const bidirectional iterator for the basic_json class /// a const bidirectional iterator for the basic_json class
...@@ -2425,49 +2409,68 @@ class basic_json ...@@ -2425,49 +2409,68 @@ class basic_json
/// associated JSON instance /// associated JSON instance
pointer m_object = nullptr; pointer m_object = nullptr;
/// the actual iterator of the associated instance /// the actual iterator of the associated instance
internal_const_iterator m_it; internal_iterator<typename array_t::const_iterator, typename object_t::const_iterator> m_it;
}; };
private: private:
//////////// //////////////////////
// parser // // lexer and parser //
//////////// //////////////////////
/*!
@brief lexical analysis
This class organizes the lexical analysis during JSON deserialization. The
core of it is a scanner generated by re2c <http://re2c.org> that processes
a buffer and recognizes tokens according to RFC 7159 and ECMA-404.
*/
class lexer class lexer
{ {
public: public:
/// token types for the parser /// token types for the parser
enum class token_type enum class token_type
{ {
uninitialized, uninitialized, ///< indicating the scanner is uninitialized
literal_true, literal_true, ///< the "true" literal
literal_false, literal_false, ///< the "false" literal
literal_null, literal_null, ///< the "null" literal
value_string, value_string, ///< a string - use get_string() for actual value
value_number, value_number, ///< a number - use get_number() for actual value
begin_array, begin_array, ///< the character for array begin "["
begin_object, begin_object, ///< the character for object begin "{"
end_array, end_array, ///< the character for array end "]"
end_object, end_object, ///< the character for object end "}"
name_separator, name_separator, ///< the name separator ":"
value_separator, value_separator, ///< the value separator ","
parse_error, parse_error, ///< indicating a parse error
end_of_input end_of_input ///< indicating the end of the input buffer
}; };
/// the char type to use in the lexer /// the char type to use in the lexer
using lexer_char_t = typename string_t::value_type; using lexer_char_t = typename string_t::value_type;
inline lexer(const typename string_t::value_type* s) : m_content(s) /// constructor with a given buffer
inline lexer(const string_t& s) noexcept
: m_content(s.c_str())
{ {
m_start = m_cursor = m_content; m_start = m_cursor = m_content;
m_limit = m_content + strlen(m_content); m_limit = m_content + s.size();
} }
/// default constructor
inline lexer() = default; inline lexer() = default;
inline static string_t to_unicode(const long codepoint) /*!
@brief create a string from a Unicode code point
@param codepoint the code point (must be in [0x0, 0x10ffff]
@return string representation of the code point
@exception std::out_of_range if code point is >0x10ffff
@see <http://en.wikipedia.org/wiki/UTF-8#Sample_code>
*/
inline static string_t to_unicode(const size_t codepoint)
{ {
string_t result; string_t result;
...@@ -2505,7 +2508,8 @@ class basic_json ...@@ -2505,7 +2508,8 @@ class basic_json
return result; return result;
} }
inline static std::string token_type_name(token_type t) /// return name of values of type token_type
inline static std::string token_type_name(token_type t) noexcept
{ {
switch (t) switch (t)
{ {
...@@ -2544,16 +2548,16 @@ class basic_json ...@@ -2544,16 +2548,16 @@ class basic_json
This function implements a scanner for JSON. It is specified using This function implements a scanner for JSON. It is specified using
regular expressions that try to follow RFC 7159 and ECMA-404 as close regular expressions that try to follow RFC 7159 and ECMA-404 as close
as possible. These regular expressions are then translated into a as possible. These regular expressions are then translated into a
deterministic finite automaton (DFA) by the tool RE2C. As a result, the deterministic finite automaton (DFA) by the tool re2c
translated code for this function consists of a large block of code <http://re2c.org>. As a result, the translated code for this function
with goto jumps. consists of a large block of code with goto jumps.
@return the class of the next token read from the buffer @return the class of the next token read from the buffer
*/ */
inline token_type scan() inline token_type scan() noexcept
{ {
// pointer for backtracking information // pointer for backtracking information
const typename string_t::value_type* m_marker = nullptr; const lexer_char_t* m_marker = nullptr;
// remember the begin of the token // remember the begin of the token
m_start = m_cursor; m_start = m_cursor;
...@@ -2618,19 +2622,32 @@ class basic_json ...@@ -2618,19 +2622,32 @@ class basic_json
*/ */
} }
inline string_t get_token() const /// return string representation of last read token
inline string_t get_token() const noexcept
{ {
return string_t(m_start, static_cast<size_t>(m_cursor - m_start)); return string_t(m_start, static_cast<size_t>(m_cursor - m_start));
} }
/*! /*!
The pointer m_start points to the opening quote of the string, and @brief return string value for string tokens
m_cursor past the closing quote of the string. We create a std::string
from the character after the opening quotes (m_start+1) until the The function iterates the characters between the opening and closing
character before the closing quotes (hence subtracting 2 characters quotes of the string value. The complete string is the range
from the pointer difference of the two pointers). [m_start,m_cursor). Consequently, we iterate from m_start+1 to
m_cursor-1.
We differentiate two cases:
1. Escaped characters. In this case, a new character is constructed
according to the nature of the escape. Some escapes create new
characters (e.g., @c "\\n" is replaced by @c "\n"), some are copied
as is (e.g., @c "\\\\"). Furthermore, Unicode escapes of the shape
@c "\\uxxxx" need special care. In this case, to_unicode takes care
of the construction of the values.
2. Unescaped characters are copied as is.
@return string value of current token without opening and closing quotes @return string value of current token without opening and closing quotes
@exception std::out_of_range if to_unicode fails
*/ */
inline string_t get_string() const inline string_t get_string() const
{ {
...@@ -2696,7 +2713,7 @@ class basic_json ...@@ -2696,7 +2713,7 @@ class basic_json
case 'u': case 'u':
{ {
// get code xxxx from \uxxxx // get code xxxx from \uxxxx
auto codepoint = strtol(i + 1, nullptr, 16); auto codepoint = std::strtoul(i + 1, nullptr, 16);
// add unicode character(s) // add unicode character(s)
result += to_unicode(codepoint); result += to_unicode(codepoint);
// skip the next four characters (\uxxxx) // skip the next four characters (\uxxxx)
...@@ -2716,19 +2733,29 @@ class basic_json ...@@ -2716,19 +2733,29 @@ class basic_json
return result; return result;
} }
/*!
@brief return number value for number tokens
This function translates the last token into a floating point number.
The pointer m_begin points to the beginning of the parsed number. We
pass this pointer to std::strtod which sets endptr to the first
character past the converted number. If this pointer is not the same as
m_cursor, then either more or less characters have been used during the
comparison. This can happen for inputs like "01" which will be treated
like number 0 followed by number 1.
@return the result of the number conversion or NAN if the conversion
read past the current token. The latter case needs to be treated by the
caller function.
@exception std::range_error if passed value is out of range
*/
inline number_float_t get_number() const inline number_float_t get_number() const
{ {
// The pointer m_begin points to the beginning of the parsed
// number. We pass this pointer to std::strtod which sets endptr to
// the first character past the converted number. If this pointer is
// not the same as m_cursor, then either more or less characters
// have been used during the comparison. This can happen for inputs
// like "01" which will be treated like number 0 followed by number
// 1.
// conversion // conversion
char* endptr; lexer_char_t* endptr;
const auto float_val = std::strtod(reinterpret_cast<const char*>(m_start), &endptr); const auto float_val = std::strtod(reinterpret_cast<const lexer_char_t*>(m_start),
&endptr);
// return float_val if the whole number was translated and NAN // return float_val if the whole number was translated and NAN
// otherwise // otherwise
...@@ -2737,20 +2764,20 @@ class basic_json ...@@ -2737,20 +2764,20 @@ class basic_json
private: private:
/// the buffer /// the buffer
const typename string_t::value_type* m_content = nullptr; const lexer_char_t* m_content = nullptr;
/// pointer to he beginning of the current symbol /// pointer to he beginning of the current symbol
const typename string_t::value_type* m_start = nullptr; const lexer_char_t* m_start = nullptr;
/// pointer to the current symbol /// pointer to the current symbol
const typename string_t::value_type* m_cursor = nullptr; const lexer_char_t* m_cursor = nullptr;
/// pointer to the end of the buffer /// pointer to the end of the buffer
const typename string_t::value_type* m_limit = nullptr; const lexer_char_t* m_limit = nullptr;
}; };
class parser class parser
{ {
public: public:
/// constructor for strings /// constructor for strings
inline parser(const string_t& s) : m_buffer(s), m_lexer(m_buffer.c_str()) inline parser(const string_t& s) : m_buffer(s), m_lexer(m_buffer)
{ {
// read first token // read first token
get_token(); get_token();
...@@ -2767,7 +2794,7 @@ class basic_json ...@@ -2767,7 +2794,7 @@ class basic_json
} }
// initializer lexer // initializer lexer
m_lexer = lexer(m_buffer.c_str()); m_lexer = lexer(m_buffer);
// read first token // read first token
get_token(); get_token();
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment