Namespaces
Variants
Views
Actions

Difference between revisions of "cpp/utility/format/range formatter"

From cppreference.com
< cpp‎ | utility‎ | format
(added c++23 std::range_formatter)
 
m (`assert` is more expressive.)
 
(11 intermediate revisions by 4 users not shown)
Line 6: Line 6:
 
{{dcl|since=c++23|1=
 
{{dcl|since=c++23|1=
 
template< class T, class CharT = char >
 
template< class T, class CharT = char >
−
  requires std::same_as<std::remove_cvref_t<T>, T> && std::formattable<T, CharT>
+
    requires std::same_as<std::remove_cvref_t<T>, T> && std::formattable<T, CharT>
 
class range_formatter;
 
class range_formatter;
 
}}
 
}}
Line 21: Line 21:
 
{{sdsc end}}
 
{{sdsc end}}
  
−
The {{spar|range-fill-and-align}} is interpreted the same way as a {{spar|fill-and-align}} except that the {{spar|fill}} in {{spar|range-fill-and-align}} is any character other than {{ttb|<nowiki>{</nowiki>}}, {{ttb|<nowiki>}</nowiki>}}, or {{ttb|:}}.  
+
The {{spar|range-fill-and-align}} is interpreted the same way as a {{spar|fill-and-align}} except that the {{spar|fill}} in {{spar|range-fill-and-align}} is any character other than {{ttb|{}}, {{ttb|}<!---->}}, or {{ttb|:}}.
  
−
The {{spar|width}} is described in [[cpp/utility/format/formatter#Width and precision|standard format width specification]].
+
The {{spar|width}} is described in {{rlp|spec#Width and precision|standard format width specification}}.
  
 
The {{ttb|n}} option causes the range to be formatted without the opening and closing brackets.
 
The {{ttb|n}} option causes the range to be formatted without the opening and closing brackets.
 +
{{source|1=
 +
assert(std::format("{}", views::iota(1, 5)) == "[1, 2, 3, 4]");
 +
assert(std::format("{:n}", views::iota(1, 5)) == "1, 2, 3, 4");
 +
}}
  
 
The {{spar|format-spec}} in a {{spar|range-underlying-spec}} (its syntax is equivalent to {{tt|:}} {{spar|format-spec}}), if any, is interpreted by the range element formatter {{tt|std::formatter<T, CharT>}}.
 
The {{spar|format-spec}} in a {{spar|range-underlying-spec}} (its syntax is equivalent to {{tt|:}} {{spar|format-spec}}), if any, is interpreted by the range element formatter {{tt|std::formatter<T, CharT>}}.
 +
{{source|1=
 +
std::array ints{12, 10, 15, 14};
  
−
The {{spar|range-type}} changes the way a range is formatted, with certain options only valid with certain argument types.  
+
assert(std::format("{}", ints) == "[12, 10, 15, 14]");
 +
assert(std::format("{::X}", ints) == "[C, A, F, E]");
 +
assert(std::format("{:n:_^4}", ints) == "_12_, _10_, _15_, _14_");
 +
}}
 +
 
 +
The {{spar|range-type}} changes the way a range is formatted, with certain options only valid with certain argument types.
  
 
The available range presentation types are:
 
The available range presentation types are:
−
* {{ttb|m}}: Indicates that the opening bracket should be {{c|"{"}}, the closing bracket should be {{c|"}"}}, the separator should be {{c|", "}}, and each range element should be formatted as if {{ttb|m}} were specified for its {{spar|tuple-type}} (from {{spar|tuple-format-spec}}).
+
* {{ttb|m}}: Indicates that the opening bracket should be {{c|"{"}}, the closing bracket should be {{c|"}"}}, the separator should be {{c|", "}}, and each range element should be formatted as if {{ttb|m}} were specified for its {{spar|tuple-type}} (in [[cpp/utility/format/tuple_formatter#Format specification|{{spar|tuple-format-spec}}]]).
  
 
:* If {{ttb|m}} is chosen as the {{spar|range-type}}, the program is ill-formed unless {{tt|T}} is either a specialization of:
 
:* If {{ttb|m}} is chosen as the {{spar|range-type}}, the program is ill-formed unless {{tt|T}} is either a specialization of:
 
::* {{lc|std::pair}}, or
 
::* {{lc|std::pair}}, or
−
::* {{lc|std::tuple}} such that {{c|std::tuple_size_v<T> {{==}} 2}} is {{c|true}}.
+
::* {{lc|std::tuple}} such that {{c|1=std::tuple_size_v<T> == 2}} is {{c|true}}.
 +
{{source|1=
 +
std::array char_pairs
 +
{
 +
    std::pair{'A', 5}, std::pair{'B', 10}, std::pair{'C', 12}
 +
};
  
−
* {{ttb|s}}: Indicates that the range should be formatted as a string.
+
assert(std::format("{}", char_pairs) == "[('A', 5), ('B', 10), ('C', 12)]");
−
* {{ttb|?s}}: Indicates that the range should be formatted as an [[cpp/utility/format/formatter#Formatting_escaped_characters_and_strings|escaped string]].
+
assert(std::format("{:m}", char_pairs) == "{'A': 5, 'B': 10, 'C': 12}");
 +
}}
  
 +
* {{ttb|s}}: Indicates that the range should be formatted as a string.
 +
* {{ttb|?s}}: Indicates that the range should be formatted as an {{rlp|spec#Formatting escaped characters and strings|escaped string}}.
 
:* If {{ttb|s}} or {{ttb|?s}} is chosen as the {{spar|range-type}}, both {{ttb|n}} option and {{spar|range-underlying-spec}} should not be included in the format specifier, and
 
:* If {{ttb|s}} or {{ttb|?s}} is chosen as the {{spar|range-type}}, both {{ttb|n}} option and {{spar|range-underlying-spec}} should not be included in the format specifier, and
 
:* the program is ill-formed unless {{tt|T}} is {{tt|CharT}}.
 
:* the program is ill-formed unless {{tt|T}} is {{tt|CharT}}.
 +
{{source|1=
 +
std::array star{'S', 'T', 'A', 'R'};
 +
 +
assert(std::format("{}", star) == "['S', 'T', 'A', 'R']");
 +
assert(std::format("{:s}", star) == "STAR");
 +
assert(std::format("{:?s}", star) == "\"STAR\"");
 +
}}
  
 
===Member objects===
 
===Member objects===
 
{{dsc begin}}
 
{{dsc begin}}
 
{{dsc hitem|Member name|Definition}}
 
{{dsc hitem|Member name|Definition}}
−
{{dsc|{{tti|underlying_}} {{mark|private}}|the underlying formatter<br>{{mark expos mem obj}}}}
+
{{dsc expos mem obj|private=yes|underlying_|the underlying formatter of type {{c/core|std::formatter<T, CharT>}}}}
−
{{dsc|{{tti|separator_}} {{mark|private}}|a string representing the separator of the range formatted result<br>{{mark expos mem obj}}}}
+
{{dsc expos mem obj|private=yes|separator_|a string representing the separator of the range formatted result. The default separator is {{c|", "}}.}}
−
{{dsc|{{tti|opening-bracket_}} {{mark|private}}|a string representing the opening bracket of the range formatted result<br>{{mark expos mem obj}}}}
+
{{dsc expos mem obj|private=yes|opening-bracket_|a string representing the opening bracket of the range formatted result. The default opening bracket is {{c|"["}}.}}
−
{{dsc|{{tti|closing-bracket_}} {{mark|private}}|a string representing the closing bracket of the range formatted result<br>{{mark expos mem obj}}}}
+
{{dsc expos mem obj|private=yes|closing-bracket_|a string representing the closing bracket of the range formatted result. The default closing bracket is {{c|"]"}}.}}
 
{{dsc end}}
 
{{dsc end}}
  
Line 62: Line 88:
 
{{dsc end}}
 
{{dsc end}}
  
−
{{member | 1={{small|std::range_formatter::}}set_separator |2=
+
{{member|1={{small|std::range_formatter::}}set_separator|2=
 
{{dcl begin}}
 
{{dcl begin}}
−
{{dcl | 1=
+
{{dcl|1=
 
constexpr void set_separator( std::basic_string_view<CharT> sep ) noexcept;
 
constexpr void set_separator( std::basic_string_view<CharT> sep ) noexcept;
 
}}
 
}}
 
{{dcl end}}
 
{{dcl end}}
  
−
Assigns {{tt|sep}} to {{tti|separator_}}.
+
Assigns {{c|sep}} to {{tti|separator_}}.
 
}}
 
}}
  
−
{{member | 1={{small|std::range_formatter::}}set_brackets |2=
+
{{member|1={{small|std::range_formatter::}}set_brackets|2=
 
{{dcl begin}}
 
{{dcl begin}}
−
{{dcl | 1=
+
{{dcl|1=
 
constexpr void set_brackets( std::basic_string_view<CharT> opening,
 
constexpr void set_brackets( std::basic_string_view<CharT> opening,
 
                             std::basic_string_view<CharT> closing ) noexcept;
 
                             std::basic_string_view<CharT> closing ) noexcept;
Line 80: Line 106:
 
{{dcl end}}
 
{{dcl end}}
  
−
Assigns {{tt|opening}} and {{tt|closing}} to {{tti|opening-bracket_}} and {{tti|closing-bracket_}}, respectively.
+
Assigns {{c|opening}} and {{c|closing}} to {{tti|opening-bracket_}} and {{tti|closing-bracket_}}, respectively.
 
}}
 
}}
  
−
{{member | 1={{small|std::range_formatter::}}parse |2=
+
{{member|1={{small|std::range_formatter::}}underlying|2=
 
{{dcl begin}}
 
{{dcl begin}}
−
{{dcl | 1=
+
{{dcl|num=1|1=
 +
constexpr std::formatter<T, CharT>& underlying();
 +
}}
 +
{{dcl|num=2|1=
 +
constexpr const std::formatter<T, CharT>& underlying() const;
 +
}}
 +
{{dcl end}}
 +
 
 +
Returns {{box|{{tti|underlying_}}}} (the underlying formatter).
 +
}}
 +
 
 +
{{member|1={{small|std::range_formatter::}}parse|2=
 +
{{dcl begin}}
 +
{{dcl|1=
 
template< class ParseContext >
 
template< class ParseContext >
−
constexpr auto parse( ParseContext& ctx ) -> typename ParseContext::iterator;
+
constexpr auto parse( ParseContext& ctx ) -> ParseContext::iterator;
 
}}
 
}}
 
{{dcl end}}
 
{{dcl end}}
  
−
Parses the format specifier as a {{spar|range-format-spec}} and stores the parsed specifiers in the current object.
+
Parses the format specifiers as a {{spar|range-format-spec}} and stores the parsed specifiers in the current object.
  
−
If {{spar|range-type}} or the {{ttb|n}} option is present, the values of {{tti|opening-bracket}}, {{tti|closing-bracket}}, and {{tti|separator}} are modified as required.
+
Calls {{box|{{tti|underlying_}}{{c/core|.parse(ctx)}}}} to parse {{spar|format-spec}} in {{spar|range-format-spec}} or, if the latter is not present, an empty {{spar|format-spec}}.
  
−
It calls {{c|underlying_.set_debug_format()}} if:
+
If {{spar|range-type}} or the {{ttb|n}} option is present, the values of {{tti|opening-bracket_}}, {{tti|closing-bracket_}}, and {{tti|separator_}} are modified as required.
 +
 
 +
It calls {{box|{{tti|underlying_}}{{c/core|.set_debug_format()}}}} if:
 
* the {{spar|range-type}} is neither {{ttb|s}} nor {{ttb|?s}},
 
* the {{spar|range-type}} is neither {{ttb|s}} nor {{ttb|?s}},
−
* {{c|underlying_.set_debug_format()}} is a valid expression, and
+
* {{box|{{tti|underlying_}}{{c/core|.set_debug_format()}}}} is a valid expression, and
 
* there is no {{spar|range-underlying-spec}}.
 
* there is no {{spar|range-underlying-spec}}.
  
Line 103: Line 144:
 
}}
 
}}
  
−
{{member | 1={{small|std::range_formatter::}}format |2=
+
{{member|1={{small|std::range_formatter::}}format|2=
 
{{dcl begin}}
 
{{dcl begin}}
−
{{dcl | 1=
+
{{dcl|1=
 
template< ranges::input_range R, class FormatContext >
 
template< ranges::input_range R, class FormatContext >
 
   requires std::formattable<ranges::range_reference_t<R>, CharT> &&
 
   requires std::formattable<ranges::range_reference_t<R>, CharT> &&
 
           std::same_as<std::remove_cvref_t<ranges::range_reference_t<R>>, T>
 
           std::same_as<std::remove_cvref_t<ranges::range_reference_t<R>>, T>
−
auto format( R&& r, FormatContext& ctx ) const -> typename FormatContext::iterator;
+
auto format( R&& r, FormatContext& ctx ) const -> FormatContext::iterator;
 
}}
 
}}
 
{{dcl end}}
 
{{dcl end}}
Line 116: Line 157:
  
 
Otherwise, it writes the following into {{c|ctx.out()}} as specified by {{spar|range-format-spec}}, in order:
 
Otherwise, it writes the following into {{c|ctx.out()}} as specified by {{spar|range-format-spec}}, in order:
−
* {{tti|opening-bracket_}},
+
* {{box|{{tti|opening-bracket_}}}},
−
* for each formattable element {{tt|e}} of the range {{tt|r}}:
+
* for each formattable element {{c|e}} of the range {{c|r}}:
−
:* the result of writing {{tt|e}} view {{tti|underlying_}}, and
+
:* the result of writing {{c|e}} via {{tti|underlying_}}, and
−
:* {{tti|separator_}}, unless {{tt|e}} is the last element of {{tt|r}}, and
+
:* {{box|{{tti|separator_}}}}, unless {{c|e}} is the last element of {{c|r}}, and
−
* {{tti|opening-bracket_}}.
+
* {{box|{{tti|closing-bracket_}}}}.
  
 
Returns an iterator past the end of the output range.
 
Returns an iterator past the end of the output range.
 
}}
 
}}
 +
 +
===Defect reports===
 +
{{dr list begin}}
 +
{{dr list item|wg=lwg|dr=3892|std=c++23|before=the formatting of nested ranges was incorrect|after=corrected}}
 +
{{dr list end}}
  
 
===See also===
 
===See also===
Line 129: Line 175:
 
{{dsc inc|cpp/utility/format/dsc formatter}}
 
{{dsc inc|cpp/utility/format/dsc formatter}}
 
{{dsc end}}
 
{{dsc end}}
 +
 +
{{langlinks|es|ja|ru|zh}}

Latest revision as of 14:21, 11 September 2024

 
C++
 
Utilities library
General utilities
Relational operators (deprecated in C++20)
 
Formatting library
Standard format specification
Formatting functions
(C++20)
(C++20)
(C++20)
(C++20)
(C++20)
(C++20)
Format strings
(C++20)(C++20)(C++20)
(C++26)
Formatting concepts
(C++23)
Formatter
(C++20)
(C++23)
(C++23)
range_formatter
(C++23)
(C++23)
(C++20)(C++20)(C++20)
(C++20)(C++20)(C++20)
(C++23)
(C++23)
Formatting arguments
(C++20)
(C++20)
(C++20)(C++20)(C++20)
(C++20) (deprecated in C++26)
(C++20)(C++20)
Format error
(C++20)
 
Defined in header <format>
template< class T, class CharT = char >

    requires std::same_as<std::remove_cvref_t<T>, T> && std::formattable<T, CharT>

class range_formatter;
 (since C++23)

The std::range_formatter is a helper class template for implementing std::formatter specializations for range types.

Contents

[edit] Range format specification

The syntax of range-format-spec is:

range-fill-and-align (optional) width (optional) n(optional) range-type (optional) range-underlying-spec (optional)

The range-fill-and-align is interpreted the same way as a fill-and-align except that the fill in range-fill-and-align is any character other than {, }, or :.

The width is described in standard format width specification.

The n option causes the range to be formatted without the opening and closing brackets. 

assert(std::format("{}", views::iota(1, 5)) == "[1, 2, 3, 4]");
assert(std::format("{:n}", views::iota(1, 5)) == "1, 2, 3, 4");

The format-spec in a range-underlying-spec (its syntax is equivalent to : format-spec), if any, is interpreted by the range element formatter std::formatter<T, CharT>. 

std::array ints{12, 10, 15, 14};
 
assert(std::format("{}", ints) == "[12, 10, 15, 14]");
assert(std::format("{::X}", ints) == "[C, A, F, E]");
assert(std::format("{:n:_^4}", ints) == "_12_, _10_, _15_, _14_");

The range-type changes the way a range is formatted, with certain options only valid with certain argument types.

The available range presentation types are:

  • m: Indicates that the opening bracket should be "{", the closing bracket should be "}", the separator should be ", ", and each range element should be formatted as if m were specified for its tuple-type (in tuple-format-spec).
  • If m is chosen as the range-type, the program is ill-formed unless T is either a specialization of: 
std::array char_pairs
{
    std::pair{'A', 5}, std::pair{'B', 10}, std::pair{'C', 12}
};
 
assert(std::format("{}", char_pairs) == "[('A', 5), ('B', 10), ('C', 12)]");
assert(std::format("{:m}", char_pairs) == "{'A': 5, 'B': 10, 'C': 12}");
  • s: Indicates that the range should be formatted as a string.
  • ?s: Indicates that the range should be formatted as an escaped string.
  • If s or ?s is chosen as the range-type, both n option and range-underlying-spec should not be included in the format specifier, and
  • the program is ill-formed unless T is CharT. 
std::array star{'S', 'T', 'A', 'R'};
 
assert(std::format("{}", star) == "['S', 'T', 'A', 'R']");
assert(std::format("{:s}", star) == "STAR");
assert(std::format("{:?s}", star) == "\"STAR\"");

[edit] Member objects

Member name Definition
underlying_ (private) the underlying formatter of type std::formatter<T, CharT>
(exposition-only member object*)
separator_ (private) a string representing the separator of the range formatted result. The default separator is ", ".
(exposition-only member object*)
opening-bracket_ (private) a string representing the opening bracket of the range formatted result. The default opening bracket is "[".
(exposition-only member object*)
closing-bracket_ (private) a string representing the closing bracket of the range formatted result. The default closing bracket is "]".
(exposition-only member object*)

[edit] Member functions

set_separator
 sets a specified separator for the range formatted result
(public member function)
set_brackets
 sets a specified opening and closing brackets for the range formatted result
(public member function)
underlying
 returns the underlying formatter
(public member function)
parse
 parses the format specifier as specified by range-format-spec
(public member function)
format
 writes the range formatted output as specified by range-format-spec
(public member function)

std::range_formatter::set_separator

constexpr void set_separator( std::basic_string_view<CharT> sep ) noexcept;

Assigns sep to separator_.

std::range_formatter::set_brackets

constexpr void set_brackets( std::basic_string_view<CharT> opening,
                             std::basic_string_view<CharT> closing ) noexcept;

Assigns opening and closing to opening-bracket_ and closing-bracket_, respectively.

std::range_formatter::underlying

constexpr std::formatter<T, CharT>& underlying();
 (1)
constexpr const std::formatter<T, CharT>& underlying() const;
 (2)

Returns underlying_ (the underlying formatter).

std::range_formatter::parse

template< class ParseContext >
constexpr auto parse( ParseContext& ctx ) -> ParseContext::iterator;

Parses the format specifiers as a range-format-spec and stores the parsed specifiers in the current object.

Calls underlying_.parse(ctx) to parse format-spec in range-format-spec or, if the latter is not present, an empty format-spec.

If range-type or the n option is present, the values of opening-bracket_, closing-bracket_, and separator_ are modified as required.

It calls underlying_.set_debug_format() if:

  • the range-type is neither s nor ?s,
  • underlying_.set_debug_format() is a valid expression, and
  • there is no range-underlying-spec.

Returns an iterator past the end of the range-format-spec.

std::range_formatter::format

template< ranges::input_range R, class FormatContext >

  requires std::formattable<ranges::range_reference_t<R>, CharT> &&
           std::same_as<std::remove_cvref_t<ranges::range_reference_t<R>>, T>

auto format( R&& r, FormatContext& ctx ) const -> FormatContext::iterator;

If the range-type was either s or ?s, it writes the formatted std::basic_string<CharT>(std::from_range, r) as a string or an escaped string, respectively, into ctx.out().

Otherwise, it writes the following into ctx.out() as specified by range-format-spec, in order:

  • opening-bracket_,
  • for each formattable element e of the range r:
  • the result of writing e via underlying_, and
  • separator_, unless e is the last element of r, and
  • closing-bracket_.

Returns an iterator past the end of the output range.

[edit] Defect reports

The following behavior-changing defect reports were applied retroactively to previously published C++ standards.

DR Applied to Behavior as published Correct behavior
LWG 3892 C++23 the formatting of nested ranges was incorrect corrected

[edit] See also

(C++20)
 defines formatting rules for a given type
(class template) [edit]
Retrieved from "https://en.cppreference.com/mwiki/index.php?title=cpp/utility/format/range_formatter&oldid=176024"