hoshi-lang dev
Yet another programming language
Loading...
Searching...
No Matches
lexer< BasicJsonType, InputAdapterType > Class Template Reference

lexical analysis More...

Inheritance diagram for lexer< BasicJsonType, InputAdapterType >:
Collaboration diagram for lexer< BasicJsonType, InputAdapterType >:

Public Types

using token_type = typename lexer_base< BasicJsonType >::token_type
 
- Public Types inherited from lexer_base< BasicJsonType >
enum class  token_type {
  uninitialized , literal_true , literal_false , literal_null ,
  value_string , value_unsigned , value_integer , value_float ,
  begin_array , begin_object , end_array , end_object ,
  name_separator , value_separator , parse_error , end_of_input ,
  literal_or_value
}
 token types for the parser More...
 

Public Member Functions

 lexer (InputAdapterType &&adapter, bool ignore_comments_=false) noexcept
 
 lexer (const lexer &)=delete
 
 lexer (lexer &&)=default
 
lexeroperator= (lexer &)=delete
 
lexeroperator= (lexer &&)=default
 
 ~lexer ()=default
 
constexpr number_integer_t get_number_integer () const noexcept
 return integer value
 
constexpr number_unsigned_t get_number_unsigned () const noexcept
 return unsigned integer value
 
constexpr number_float_t get_number_float () const noexcept
 return floating-point value
 
string_tget_string ()
 return current string value (implicitly resets the token; useful only once)
 
constexpr position_t get_position () const noexcept
 return position of last read token
 
std::string get_token_string () const
 
JSON_HEDLEY_RETURNS_NON_NULL constexpr const char * get_error_message () const noexcept
 return syntax error message
 
bool skip_bom ()
 skip the UTF-8 byte order mark
 
void skip_whitespace ()
 
token_type scan ()
 

Private Types

using number_integer_t = typename BasicJsonType::number_integer_t
 
using number_unsigned_t = typename BasicJsonType::number_unsigned_t
 
using number_float_t = typename BasicJsonType::number_float_t
 
using string_t = typename BasicJsonType::string_t
 
using char_type = typename InputAdapterType::char_type
 
using char_int_type = typename char_traits< char_type >::int_type
 

Private Member Functions

int get_codepoint ()
 get codepoint from 4 hex characters following \u
 
bool next_byte_in_range (std::initializer_list< char_int_type > ranges)
 check if the next byte(s) are inside a given range
 
token_type scan_string ()
 scan a string literal
 
bool scan_comment ()
 scan a comment
 
token_type scan_number ()
 scan a number literal
 
token_type scan_literal (const char_type *literal_text, const std::size_t length, token_type return_type)
 
void reset () noexcept
 reset token_buffer; current character is beginning of token
 
char_int_type get ()
 
void unget ()
 unget current character (read it again on next get)
 
void add (char_int_type c)
 add a character to token_buffer
 

Static Private Member Functions

static JSON_HEDLEY_PURE char get_decimal_point () noexcept
 return the locale-dependent decimal point
 
static void strtof (float &f, const char *str, char **endptr) noexcept
 
static void strtof (double &f, const char *str, char **endptr) noexcept
 
static void strtof (long double &f, const char *str, char **endptr) noexcept
 

Private Attributes

InputAdapterType ia
 input adapter
 
const bool ignore_comments = false
 whether comments should be ignored (true) or signaled as errors (false)
 
char_int_type current = char_traits<char_type>::eof()
 the current character
 
bool next_unget = false
 whether the next get() call should just return current
 
position_t position {}
 the start position of the current token
 
std::vector< char_typetoken_string {}
 raw input token string (for error messages)
 
string_t token_buffer {}
 buffer for variable-length tokens (numbers, strings)
 
const char * error_message = ""
 a description of occurred lexer errors
 
number_integer_t value_integer = 0
 
number_unsigned_t value_unsigned = 0
 
number_float_t value_float = 0
 
const char_int_type decimal_point_char = '.'
 the decimal point
 
std::size_t decimal_point_position = std::string::npos
 the position of the decimal point in the input
 

Additional Inherited Members

- Static Public Member Functions inherited from lexer_base< BasicJsonType >
JSON_HEDLEY_RETURNS_NON_NULL static JSON_HEDLEY_CONST const char * token_type_name (const token_type t) noexcept
 return name of values of type token_type (only used for errors)
 

Detailed Description

template<typename BasicJsonType, typename InputAdapterType>
class detail::lexer< BasicJsonType, InputAdapterType >

lexical analysis

This class organizes the lexical analysis during JSON deserialization.

Definition at line 7134 of file json.hpp.

Member Typedef Documentation

◆ char_int_type

template<typename BasicJsonType , typename InputAdapterType >
using char_int_type = typename char_traits<char_type>::int_type
private

Definition at line 7141 of file json.hpp.

◆ char_type

template<typename BasicJsonType , typename InputAdapterType >
using char_type = typename InputAdapterType::char_type
private

Definition at line 7140 of file json.hpp.

◆ number_float_t

template<typename BasicJsonType , typename InputAdapterType >
using number_float_t = typename BasicJsonType::number_float_t
private

Definition at line 7138 of file json.hpp.

◆ number_integer_t

template<typename BasicJsonType , typename InputAdapterType >
using number_integer_t = typename BasicJsonType::number_integer_t
private

Definition at line 7136 of file json.hpp.

◆ number_unsigned_t

template<typename BasicJsonType , typename InputAdapterType >
using number_unsigned_t = typename BasicJsonType::number_unsigned_t
private

Definition at line 7137 of file json.hpp.

◆ string_t

template<typename BasicJsonType , typename InputAdapterType >
using string_t = typename BasicJsonType::string_t
private

Definition at line 7139 of file json.hpp.

◆ token_type

template<typename BasicJsonType , typename InputAdapterType >
using token_type = typename lexer_base<BasicJsonType>::token_type

Definition at line 7144 of file json.hpp.

Constructor & Destructor Documentation

◆ lexer() [1/3]

template<typename BasicJsonType , typename InputAdapterType >
lexer ( InputAdapterType &&  adapter,
bool  ignore_comments_ = false 
)
inlineexplicitnoexcept

Definition at line 7146 of file json.hpp.

◆ lexer() [2/3]

template<typename BasicJsonType , typename InputAdapterType >
lexer ( const lexer< BasicJsonType, InputAdapterType > &  )
delete

◆ lexer() [3/3]

template<typename BasicJsonType , typename InputAdapterType >
lexer ( lexer< BasicJsonType, InputAdapterType > &&  )
default

◆ ~lexer()

template<typename BasicJsonType , typename InputAdapterType >
~lexer ( )
default

Member Function Documentation

◆ add()

template<typename BasicJsonType , typename InputAdapterType >
void add ( char_int_type  c)
inlineprivate

◆ get()

◆ get_codepoint()

template<typename BasicJsonType , typename InputAdapterType >
int get_codepoint ( )
inlineprivate

get codepoint from 4 hex characters following \u

For input "\u c1 c2 c3 c4" the codepoint is: (c1 * 0x1000) + (c2 * 0x0100) + (c3 * 0x0010) + c4 = (c1 << 12) + (c2 << 8) + (c3 << 4) + (c4 << 0)

Furthermore, the possible characters '0'..'9', 'A'..'F', and 'a'..'f' must be converted to the integers 0x0..0x9, 0xA..0xF, 0xA..0xF, resp. The conversion is done by subtracting the offset (0x30, 0x37, and 0x57) between the ASCII value of the character and the desired integer value.

Returns
codepoint (0x0000..0xFFFF) or -1 in case of an error (e.g. EOF or non-hex character)

Definition at line 7192 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::current, lexer< BasicJsonType, InputAdapterType >::get(), and JSON_ASSERT.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan_string().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ get_decimal_point()

template<typename BasicJsonType , typename InputAdapterType >
static JSON_HEDLEY_PURE char get_decimal_point ( )
inlinestaticprivatenoexcept

return the locale-dependent decimal point

Definition at line 7166 of file json.hpp.

References JSON_ASSERT.

◆ get_error_message()

template<typename BasicJsonType , typename InputAdapterType >
JSON_HEDLEY_RETURNS_NON_NULL constexpr const char * get_error_message ( ) const
inlineconstexprnoexcept

return syntax error message

Definition at line 8504 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::error_message.

◆ get_number_float()

template<typename BasicJsonType , typename InputAdapterType >
constexpr number_float_t get_number_float ( ) const
inlineconstexprnoexcept

return floating-point value

Definition at line 8450 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::value_float.

◆ get_number_integer()

template<typename BasicJsonType , typename InputAdapterType >
constexpr number_integer_t get_number_integer ( ) const
inlineconstexprnoexcept

return integer value

Definition at line 8438 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::value_integer.

◆ get_number_unsigned()

template<typename BasicJsonType , typename InputAdapterType >
constexpr number_unsigned_t get_number_unsigned ( ) const
inlineconstexprnoexcept

return unsigned integer value

Definition at line 8444 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::value_unsigned.

◆ get_position()

template<typename BasicJsonType , typename InputAdapterType >
constexpr position_t get_position ( ) const
inlineconstexprnoexcept

return position of last read token

Definition at line 8471 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::position.

◆ get_string()

template<typename BasicJsonType , typename InputAdapterType >
string_t & get_string ( )
inline

return current string value (implicitly resets the token; useful only once)

Definition at line 8456 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::decimal_point_char, lexer< BasicJsonType, InputAdapterType >::decimal_point_position, and lexer< BasicJsonType, InputAdapterType >::token_buffer.

◆ get_token_string()

template<typename BasicJsonType , typename InputAdapterType >
std::string get_token_string ( ) const
inline

return the last read token (for errors only). Will never contain EOF (an arbitrary value that is not a valid char value, often -1), because 255 may legitimately occur. May contain NUL, which should be escaped.

Definition at line 8479 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::token_string.

◆ next_byte_in_range()

template<typename BasicJsonType , typename InputAdapterType >
bool next_byte_in_range ( std::initializer_list< char_int_type ranges)
inlineprivate

check if the next byte(s) are inside a given range

Adds the current byte and, for each passed range, reads a new byte and checks if it is inside the range. If a violation was detected, set up an error message and return false. Otherwise, return true.

Parameters
[in]rangeslist of integers; interpreted as list of pairs of inclusive lower and upper bound, respectively
Precondition
The passed list ranges must have 2, 4, or 6 elements; that is, 1, 2, or 3 pairs. This precondition is enforced by an assertion.
Returns
true if and only if no range violation was detected

Definition at line 7240 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::add(), lexer< BasicJsonType, InputAdapterType >::current, lexer< BasicJsonType, InputAdapterType >::error_message, lexer< BasicJsonType, InputAdapterType >::get(), JSON_ASSERT, and JSON_HEDLEY_LIKELY.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan_string().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ operator=() [1/2]

template<typename BasicJsonType , typename InputAdapterType >
lexer & operator= ( lexer< BasicJsonType, InputAdapterType > &&  )
default

◆ operator=() [2/2]

template<typename BasicJsonType , typename InputAdapterType >
lexer & operator= ( lexer< BasicJsonType, InputAdapterType > &  )
delete

◆ reset()

template<typename BasicJsonType , typename InputAdapterType >
void reset ( )
inlineprivatenoexcept

◆ scan()

◆ scan_comment()

template<typename BasicJsonType , typename InputAdapterType >
bool scan_comment ( )
inlineprivate

scan a comment

Returns
whether comment could be scanned successfully

Definition at line 7867 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::error_message, lexer< BasicJsonType, InputAdapterType >::get(), and lexer< BasicJsonType, InputAdapterType >::unget().

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ scan_literal()

template<typename BasicJsonType , typename InputAdapterType >
token_type scan_literal ( const char_type literal_text,
const std::size_t  length,
token_type  return_type 
)
inlineprivate
Parameters
[in]literal_textthe literal text to expect
[in]lengththe length of the passed literal text
[in]return_typethe token type to return on success

Definition at line 8325 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::current, lexer< BasicJsonType, InputAdapterType >::error_message, lexer< BasicJsonType, InputAdapterType >::get(), JSON_ASSERT, and JSON_HEDLEY_UNLIKELY.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ scan_number()

template<typename BasicJsonType , typename InputAdapterType >
token_type scan_number ( )
inlineprivate

scan a number literal

This function scans a string according to Sect. 6 of RFC 8259.

The function is realized with a deterministic finite state machine derived from the grammar described in RFC 8259. Starting in state "init", the input is read and used to determined the next state. Only state "done" accepts the number. State "error" is a trap state to model errors. In the table below, "anything" means any character but the ones listed before.

state 0 1-9 e E + - . anything
init zero any1 [error] [error] minus [error] [error]
minus zero any1 [error] [error] [error] [error] [error]
zero done done exponent done done decimal1 done
any1 any1 any1 exponent done done decimal1 done
decimal1 decimal2 decimal2 [error] [error] [error] [error] [error]
decimal2 decimal2 decimal2 exponent done done done done
exponent any2 any2 [error] sign sign [error] [error]
sign any2 any2 [error] [error] [error] [error] [error]
any2 any2 any2 done done done done done

The state machine is realized with one label per state (prefixed with "scan_number_") and goto statements between them. The state machine contains cycles, but any cycle can be left when EOF is read. Therefore, the function is guaranteed to terminate.

During scanning, the read bytes are stored in token_buffer. This string is then converted to a signed integer, an unsigned integer, or a floating-point number.

Returns
token_type::value_unsigned, token_type::value_integer, or token_type::value_float if number could be successfully scanned, token_type::parse_error otherwise
Note
The scanner is independent of the current locale. Internally, the locale's decimal point is used instead of . to work with the locale-dependent converters.

Definition at line 7992 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::add(), lexer< BasicJsonType, InputAdapterType >::current, lexer< BasicJsonType, InputAdapterType >::decimal_point_char, lexer< BasicJsonType, InputAdapterType >::decimal_point_position, lexer< BasicJsonType, InputAdapterType >::error_message, lexer< BasicJsonType, InputAdapterType >::get(), JSON_ASSERT, lexer< BasicJsonType, InputAdapterType >::reset(), lexer< BasicJsonType, InputAdapterType >::strtof(), lexer< BasicJsonType, InputAdapterType >::token_buffer, lexer< BasicJsonType, InputAdapterType >::unget(), lexer< BasicJsonType, InputAdapterType >::value_float, lexer< BasicJsonType, InputAdapterType >::value_integer, and lexer< BasicJsonType, InputAdapterType >::value_unsigned.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ scan_string()

template<typename BasicJsonType , typename InputAdapterType >
token_type scan_string ( )
inlineprivate

scan a string literal

This function scans a string according to Sect. 7 of RFC 8259. While scanning, bytes are escaped and copied into buffer token_buffer. Then the function returns successfully, token_buffer is not null-terminated (as it may contain \0 bytes), and token_buffer.size() is the number of bytes in the string.

Returns
token_type::value_string if string could be successfully scanned, token_type::parse_error otherwise
Note
In case of errors, variable error_message contains a textual description.

Definition at line 7277 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::add(), lexer< BasicJsonType, InputAdapterType >::current, lexer< BasicJsonType, InputAdapterType >::error_message, lexer< BasicJsonType, InputAdapterType >::get(), lexer< BasicJsonType, InputAdapterType >::get_codepoint(), JSON_ASSERT, JSON_HEDLEY_LIKELY, JSON_HEDLEY_UNLIKELY, lexer< BasicJsonType, InputAdapterType >::next_byte_in_range(), and lexer< BasicJsonType, InputAdapterType >::reset().

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ skip_bom()

template<typename BasicJsonType , typename InputAdapterType >
bool skip_bom ( )
inline

skip the UTF-8 byte order mark

Returns
true iff there is no BOM or the correct BOM has been skipped

Definition at line 8517 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::get(), and lexer< BasicJsonType, InputAdapterType >::unget().

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ skip_whitespace()

template<typename BasicJsonType , typename InputAdapterType >
void skip_whitespace ( )
inline

Definition at line 8531 of file json.hpp.

References lexer< BasicJsonType, InputAdapterType >::current, and lexer< BasicJsonType, InputAdapterType >::get().

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

Here is the call graph for this function:
Here is the caller graph for this function:

◆ strtof() [1/3]

template<typename BasicJsonType , typename InputAdapterType >
static void strtof ( double &  f,
const char *  str,
char **  endptr 
)
inlinestaticprivatenoexcept

Definition at line 7941 of file json.hpp.

◆ strtof() [2/3]

template<typename BasicJsonType , typename InputAdapterType >
static void strtof ( float &  f,
const char *  str,
char **  endptr 
)
inlinestaticprivatenoexcept

Definition at line 7935 of file json.hpp.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan_number().

Here is the caller graph for this function:

◆ strtof() [3/3]

template<typename BasicJsonType , typename InputAdapterType >
static void strtof ( long double &  f,
const char *  str,
char **  endptr 
)
inlinestaticprivatenoexcept

Definition at line 7947 of file json.hpp.

◆ unget()

template<typename BasicJsonType , typename InputAdapterType >
void unget ( )
inlineprivate

unget current character (read it again on next get)

We implement unget by setting variable next_unget to true. The input is not changed - we just simulate ungetting by modifying chars_read_total, chars_read_current_line, and token_string. The next call to get() will behave as if the unget character is read again.

Definition at line 8400 of file json.hpp.

References position_t::chars_read_current_line, position_t::chars_read_total, lexer< BasicJsonType, InputAdapterType >::current, JSON_ASSERT, JSON_HEDLEY_LIKELY, position_t::lines_read, lexer< BasicJsonType, InputAdapterType >::next_unget, lexer< BasicJsonType, InputAdapterType >::position, and lexer< BasicJsonType, InputAdapterType >::token_string.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan_comment(), lexer< BasicJsonType, InputAdapterType >::scan_number(), and lexer< BasicJsonType, InputAdapterType >::skip_bom().

Here is the caller graph for this function:

Member Data Documentation

◆ current

◆ decimal_point_char

template<typename BasicJsonType , typename InputAdapterType >
const char_int_type decimal_point_char = '.'
private

◆ decimal_point_position

template<typename BasicJsonType , typename InputAdapterType >
std::size_t decimal_point_position = std::string::npos
private

◆ error_message

◆ ia

template<typename BasicJsonType , typename InputAdapterType >
InputAdapterType ia
private

input adapter

Definition at line 8630 of file json.hpp.

Referenced by lexer< BasicJsonType, InputAdapterType >::get().

◆ ignore_comments

template<typename BasicJsonType , typename InputAdapterType >
const bool ignore_comments = false
private

whether comments should be ignored (true) or signaled as errors (false)

Definition at line 8633 of file json.hpp.

Referenced by lexer< BasicJsonType, InputAdapterType >::scan().

◆ next_unget

template<typename BasicJsonType , typename InputAdapterType >
bool next_unget = false
private

whether the next get() call should just return current

Definition at line 8639 of file json.hpp.

Referenced by lexer< BasicJsonType, InputAdapterType >::get(), and lexer< BasicJsonType, InputAdapterType >::unget().

◆ position

template<typename BasicJsonType , typename InputAdapterType >
position_t position {}
private

◆ token_buffer

template<typename BasicJsonType , typename InputAdapterType >
string_t token_buffer {}
private

◆ token_string

template<typename BasicJsonType , typename InputAdapterType >
std::vector<char_type> token_string {}
private

◆ value_float

template<typename BasicJsonType , typename InputAdapterType >
number_float_t value_float = 0
private

◆ value_integer

template<typename BasicJsonType , typename InputAdapterType >
number_integer_t value_integer = 0
private

◆ value_unsigned

template<typename BasicJsonType , typename InputAdapterType >
number_unsigned_t value_unsigned = 0
private

The documentation for this class was generated from the following file: