This file contains implementation support for the JSON path abstraction. More...

#include "sql-common/json_path.h"
#include "my_rapidjson_size_t.h"
#include <assert.h>
#include <stddef.h>
#include <algorithm>
#include <string>
#include <string_view>
#include <rapidjson/encodings.h>
#include <rapidjson/error/error.h>
#include <rapidjson/memorystream.h>
#include <rapidjson/reader.h>
#include "my_inttypes.h"
#include "mysql/strings/m_ctype.h"
#include "sql-common/json_dom.h"
#include "sql/psi_memory_key.h"
#include "sql/sql_const.h"
#include "sql_string.h"
#include "string_with_len.h"

Classes
class	anonymous_namespace{json_path.cc}::Stream
	A simple input stream class for the JSON path parser. More...

class	anonymous_namespace{json_path.cc}::MemberNameHandler
	A RapidJSON handler which accepts a scalar string and nothing else. More...

Namespaces
namespace	anonymous_namespace{json_path.cc}

Functions
static bool	is_ecmascript_identifier (const std::string_view &name)
	Returns true if the name is a valid ECMAScript identifier. More...

static bool	is_digit (unsigned codepoint)
	Return true if the codepoint is a Unicode digit. More...

static bool	is_whitespace (char ch)
	Is this a whitespace character? More...

static bool	parse_path (Stream stream, Json_path path)
	Fills in a Json_path from a path expression. More...

static bool	parse_path_leg (Stream stream, Json_path path)
	Parses a single path leg and appends it to a Json_path object. More...

static bool	parse_ellipsis_leg (Stream stream, Json_path path)
	Parses a single ellipsis leg and appends it to a Json_path object. More...

static bool	parse_array_leg (Stream stream, Json_path path)
	Parses a single array leg and appends it to a Json_path object. More...

static bool	parse_member_leg (Stream stream, Json_path path)
	Parses a single member leg and appends it to a Json_path object. More...

static bool	append_array_index (String *buf, size_t index, bool from_end)

bool	parse_path (size_t path_length, const char path_expression, Json_path path, size_t *bad_index)
	Top level parsing factory method. More...

static bool	parse_array_index (Stream stream, uint32 array_index, bool *from_end)
	Parse an array index in an array cell index or array range path leg. More...

static const char *	find_end_of_member_name (const char start, const char end)
	Find the end of a member name in a JSON path. More...

bool	parse_name_with_rapidjson (const char str, size_t len, String name)
	Parse a quoted member name using the rapidjson parser, so that we get the name without the enclosing quotes and with any escape sequences replaced with the actual characters. More...

static bool	unicode_combining_mark (unsigned codepoint)
	Return true if the character is a unicode combining mark. More...

static bool	is_letter (unsigned codepoint)
	Return true if the codepoint is a Unicode letter. More...

static bool	is_connector_punctuation (unsigned codepoint)
	Return true if the codepoint is Unicode connector punctuation. More...

bool	clone_without_autowrapping (const Json_path source_path, Json_path_clone target_path, Json_wrapper *doc, PSI_memory_key key)
	Clone a source path to a target path, stripping out legs which are made redundant by the auto-wrapping rule from the WL#7909 spec and further extended in the WL#9831 spec: More...

Variables
constexpr char	anonymous_namespace{json_path.cc}::SCOPE = '$'

constexpr char	anonymous_namespace{json_path.cc}::BEGIN_MEMBER = '.'

constexpr char	anonymous_namespace{json_path.cc}::BEGIN_ARRAY = '['

constexpr char	anonymous_namespace{json_path.cc}::END_ARRAY = ']'

constexpr char	anonymous_namespace{json_path.cc}::DOUBLE_QUOTE = '"'

constexpr char	anonymous_namespace{json_path.cc}::WILDCARD = '*'

constexpr char	anonymous_namespace{json_path.cc}::MINUS = '-'

constexpr char	anonymous_namespace{json_path.cc}::LAST [] = "last"

Detailed Description

This file contains implementation support for the JSON path abstraction.

The path abstraction is described by the functional spec attached to WL#7909.

Function Documentation

◆ append_array_index()

static bool append_array_index	(	String *	buf,
		size_t	index,
		bool	from_end
	)

static

◆ clone_without_autowrapping()

bool clone_without_autowrapping	(	const Json_path *	source_path,
		Json_path_clone *	target_path,
		Json_wrapper *	doc,
		PSI_memory_key	key
	)

Clone a source path to a target path, stripping out legs which are made redundant by the auto-wrapping rule from the WL#7909 spec and further extended in the WL#9831 spec:

"If an array cell path leg or an array range path leg is evaluated against a non-array value, the result of the evaluation is the same as if the non-array value had been wrapped in a single-element array."

See also: Json_path_leg::is_autowrap

Parameters

[in]	source_path	The original path.
[in,out]	target_path	The clone to be filled in.
[in]	doc	The document to seek through.
[in]	key	Instrumented memory key

Returns: True if an error occurred. False otherwise.

◆ find_end_of_member_name()

static const char * find_end_of_member_name	(	const char *	start,
		const char *	end
	)

static

Find the end of a member name in a JSON path.

The name could be either a quoted or an unquoted identifier.

Parameters

start	the start of the member name
end	the end of the JSON path expression

Returns: pointer to the position right after the end of the name, or to the position right after the end of the string if the input string is an unterminated quoted identifier

◆ is_connector_punctuation()

static bool is_connector_punctuation ( unsigned codepoint )

static

Return true if the codepoint is Unicode connector punctuation.

◆ is_digit()

static bool is_digit ( unsigned codepoint )

static

Return true if the codepoint is a Unicode digit.

This was the best recommendation from the old-times about how to answer this question.

◆ is_ecmascript_identifier()

static bool is_ecmascript_identifier ( const std::string_view & name )

static

Returns true if the name is a valid ECMAScript identifier.

The name must be a sequence of UTF8-encoded bytes. All escape sequences have been replaced with UTF8-encoded bytes.

Parameters

[in] name name to check

Returns: True if the name is a valid ECMAScript identifier. False otherwise.

◆ is_letter()

static bool is_letter ( unsigned codepoint )

static

Return true if the codepoint is a Unicode letter.

This was the best recommendation from the old-timers about how to answer this question. But as you can see from the need to call unicode_combining_mark(), my_isalpha() isn't good enough. It probably has many other defects.

FIXME

◆ is_whitespace()

static bool is_whitespace ( char ch )

inlinestatic

Is this a whitespace character?

◆ parse_array_index()

static bool parse_array_index	(	Stream *	stream,
		uint32 *	array_index,
		bool *	from_end
	)

static

Parse an array index in an array cell index or array range path leg.

An array index is either a non-negative integer (a 0-based index relative to the beginning of the array), or the keyword "last" (which means the last element in the array), or the keyword "last" followed by a minus ("-") and a non-negative integer (which is the 0-based index relative to the end of the array).

Parameters

[in,out]	stream	the stream to read the path expression from
[out]	array_index	gets set to the parsed array index
[out]	from_end	gets set to true if the array index is relative to the end of the array

Returns: true on error, false on success

◆ parse_array_leg()

static bool parse_array_leg	(	Stream *	stream,
		Json_path *	path
	)

static

Parses a single array leg and appends it to a Json_path object.

Parameters

[in,out]	stream	The stream to read the path expression from.
[in,out]	path	The Json_path object to fill.

Returns: true on error, false on success

◆ parse_ellipsis_leg()

static bool parse_ellipsis_leg	(	Stream *	stream,
		Json_path *	path
	)

static

Parses a single ellipsis leg and appends it to a Json_path object.

Parameters

[in,out]	stream	The stream to read the path expression from.
[in,out]	path	The Json_path object to fill.

Returns: true on error, false on success

◆ parse_member_leg()

static bool parse_member_leg	(	Stream *	stream,
		Json_path *	path
	)

static

Parses a single member leg and appends it to a Json_path object.

Parameters

[in,out]	stream	The stream to read the path expression from.
[in,out]	path	The Json_path object to fill.

Returns: true on error, false on success

◆ parse_name_with_rapidjson()

bool parse_name_with_rapidjson	(	const char *	str,
		size_t	len,
		String *	name
	)

Parse a quoted member name using the rapidjson parser, so that we get the name without the enclosing quotes and with any escape sequences replaced with the actual characters.

It is the caller's responsibility to destroy the returned Json_string when it's done with it.

Parameters

	str	the input string
	len	the length of the input string
[out]	name	the member name

Returns: false on success, true on error

◆ parse_path() [1/2]

bool parse_path	(	size_t	path_length,
		const char *	path_expression,
		Json_path *	path,
		size_t *	bad_index
	)

Top level parsing factory method.

Initialize a Json_path from a path expression.

◆ parse_path() [2/2]

static bool parse_path	(	Stream *	stream,
		Json_path *	path
	)

static

Fills in a Json_path from a path expression.

Parameters

[in,out]	stream	The stream to read the path expression from.
[in,out]	path	The Json_path object to fill.

Returns: true on error, false on success

◆ parse_path_leg()

static bool parse_path_leg	(	Stream *	stream,
		Json_path *	path
	)

static

Parses a single path leg and appends it to a Json_path object.

Parameters

[in,out]	stream	The stream to read the path expression from.
[in,out]	path	The Json_path object to fill.

Returns: true on error, false on success

◆ unicode_combining_mark()

static bool unicode_combining_mark ( unsigned codepoint )

inlinestatic

Return true if the character is a unicode combining mark.

Parameters

codepoint A unicode codepoint.

Returns: True if the codepoint is a unicode combining mark.

Classes

Namespaces

Functions

Variables

Detailed Description

Function Documentation

◆ append_array_index()

◆ clone_without_autowrapping()

◆ find_end_of_member_name()

◆ is_connector_punctuation()

◆ is_digit()

◆ is_ecmascript_identifier()

◆ is_letter()

◆ is_whitespace()

◆ parse_array_index()

◆ parse_array_leg()

◆ parse_ellipsis_leg()

◆ parse_member_leg()

◆ parse_name_with_rapidjson()

◆ parse_path() [1/2]

◆ parse_path() [2/2]

◆ parse_path_leg()

◆ unicode_combining_mark()