Document number: P0645R3
Audience: Library Evolution, Library
Author: Victor Zverovich

2018-10-07

Text Formatting

Revision History

Changes since R2

Changes since R1

Changes since R0

Introduction

Even with proliferation of graphical and voice user interfaces, text remains one of the main ways for humans to interact with computer programs and programming languages provide a variety of methods to perform text formatting. The first thing we do when learning a new programming language is often write a "Hello, World!" program that performs simple formatted output.

C++ has not one but two standard APIs for producing formatted output, the printf family of functions inherited from C and the I/O streams library (iostreams). While iostreams are usually the recommended way of producing formatted output in C++ for safety and extensibility reasons, printf offers some advantages, such as an arguably more natural function call API, the separation of formatted message and arguments, possibly with argument reordering as a POSIX extension, and often more compact source and binary code.

This paper proposes a new text formatting library that can be used as a safe and extensible alternative to the printf family of functions. It is intended to complement the existing C++ I/O streams library and reuse some of its infrastructure such as overloaded insertion operators for user-defined types.

Example:

string message = format("The answer is {}.", 42);

A full implementation of this proposal is available at https://github.com/fmtlib/fmt/tree/std.

Design

Format string syntax

Variations of the printf format string syntax are arguably the most popular among the programming languages and C++ itself inherits printf from C [1]. The advantage of the printf syntax is that many programmers are familiar with it. However, in its current form it has a number of issues:

Although it is possible to address these issues while maintaining resemblance to the original printf format, this will still break compatibility and can potentially be more confusing to users than introducing a different syntax.

Therefore we propose a new syntax based on the ones used in Python [3], the .NET family of languages [4], and Rust [5]. This syntax employs '{' and '}' as replacement field delimiters instead of '%' and it is described in detail in [format.functions]. Some advantages of the proposal are:

The syntax is expressive enough to enable translation, possibly automated, of most printf format strings. The correspondence between printf and the new syntax is given in the following table:

printfnew
-<
++
spacespace
##
00
hhunused
hunused
lunused
llunused
junused
zunused
tunused
Lunused
cc (optional)
ss (optional)
dd (optional)
id (optional)
oo
xx
XX
ud (optional)
ff
FF
ee
EE
aa
AA
gg (optional)
GG
nunused
pp (optional)

Width and precision are represented similarly in printf and the proposed syntax with the only difference that runtime value is specified by '*' in the former and '{}' in the latter, possibly with the index of the argument inside the braces:

printf("%*s", 10, "foo");
format("{:{}}", "foo", 10);

As can be seen from the table above, most of the specifiers remain the same which simplifies migration from printf. A notable difference is in the alignment specification. The proposed syntax allows left, center, and right alignment represented by '<', '^', and '>' respectively which is more expressive than the corresponding printf syntax. The latter only supports left and right alignment.

The following example uses center alignment and '*' as a fill character:

format("{:*^30}", "centered");

resulting in "***********centered***********". The same formatting cannot be easily achieved with printf.

In addition to positional arguments, the grammar can be easily extended to support named arguments.

Extensibility

Both the format string syntax and the API are designed with extensibility in mind. The mini-language can be extended for user-defined types and users can provide functions that implement parsing, possibly at compile time, and formatting for such types.

The general syntax of a replacement field in a format string is

replacement-field ::= '{' [arg-id] [':' format-spec] '}'

where format-spec is predefined for built-in types, but can be customized for user-defined types. For example, the syntax can be extended for put_time-like date and time formatting

time_t t = time(nullptr);
string date = format("The date is {0:%Y-%m-%d}.", *localtime(&t));

by providing a specialization of formatter for tm:

template<>
  struct formatter<tm> {
    constexpr parse_context::iterator parse(parse_context& ctx);

    template<class FormatContext>
      typename FormatContext::iterator format(const tm& tm, FormatContext& ctx);
  };

The formatter<tm>::parse function parses the format-spec portion of the format string corresponding to the current argument and formatter<tm>::format formats the value and writes the output via the iterator ctx.begin().

Note that date and time formatting is not covered by this proposal but formatting facilities provided by D0355 "Extending <chrono> to Calendars and Time Zones" [16] can be easily implemented using this extension API.

An implementation of formatter<T>::format can use ostream insertion operator<< for user-defined type T if available.

The extension API is based on specialization instead of the argument-dependent lookup, because the parse function doesn't take the object to be formatted as an argument and therefore some other way of parameterizing it on the argument type T such as introducing a dummy argument has to be used, e.g.

constexpr auto parse(type<T>, parse_context& ctx);

Safety

Formatting functions rely on variadic templates instead of the mechanism provided by <cstdarg>. The type information is captured automatically and passed to formatters guaranteeing type safety and making many of the printf specifiers redundant (see Format String Syntax). Memory management is automatic to prevent buffer overflow errors common to printf.

Locale support

As pointed out in P0067 "Elementary string conversions"[17] there is a number of use cases that do not require internationalization support, but do require high throughput when produced by a server. These include various text-based interchange formats such as JSON or XML. The need for locale-independent functions for conversions between integers and strings and between floating-point numbers and strings has also been highlighted in N4412: Shortcomings of iostreams. Therefore a user should be able to easily control whether to use locales or not during formatting.

We follow Python's approach [3] and designate a separate format specifier 'n' for locale-aware numeric formatting. It applies to all integral and floating-point types. All other specifiers produce output unaffected by locale settings. This can also have positive effect on performance because locale-independent formatting can be implemented more efficiently.

Positional arguments

An important feature for localization is the ability to rearrange formatting arguments as the word order may vary in different languages [7]. For example:

printf("String `%s' has %d characters\n", string, length(string)));

A possible German translation of the format string might be:

"%2$d Zeichen lang ist die Zeichenkette `%1$s'\n"

using POSIX positional arguments [2]. Unfortunately these positional specifiers are not portable [6]. The C++ I/O streams don't support such rearranging of arguments by design because they are interleaved with the portions of the literal string:

cout << "String `" << string << "' has " << length(string) << " characters\n";

The current proposal allows both positional and automatically numbered arguments, for example:

format("String `{}' has {} characters\n", string, length(string)));

with the German translation of the format string being:

"{1} Zeichen lang ist die Zeichenkette `{0}'\n"

Performance

The formatting library has been designed with performance in mind. It tries to minimize the number of virtual function calls and dynamic memory allocations done per formatting operation. In particular, if formatting output can fit into a fixed-size array allocated on stack, it should be possible to avoid both of them altogether by using a suitable API.

The format_to function takes an arbitrary output iterator and, for performance reasons, can be specialized for random-access and contiguous iterators as shown in the reference implementation [14].

The locale-independent formatting can also be implemented more efficiently than the locale-aware one. However, the main goal for the former is to support specific use cases (see Locale support) rather than to improve performance.

See Appendix A: Benchmarks for a small performance comparison of the reference implementation of this proposal versus the standard formatting facilities.

Binary footprint

In order to minimize binary code size each formatting function that uses variadic templates can be implemented as a small inline wrapper around its non-variadic counterpart. This wrapper creates a basic_format_args object, representing an array of type-erased argument references, with make_format_args and calls the non-variadic function to do the actual work. For example, the format variadic function calls vformat:

string vformat(string_view fmt, format_args args);

template<class... Args>
  inline string format(string_view fmt, const Args&... args) {
    return vformat(fmt, make_format_args(args...));
  }

basic_format_args can be implemented as an array of tagged unions. If the number of arguments is small then the tags that indicate the arguments types can be combined and passed into a formatting function as a single integer. This single integer representing all argument types is computed at compile time and only needs to be stored, resulting in smaller binary code.

Given a reasonable optimizing compiler, this will result in a compact per-call binary code, effectively consisting of placing argument pointers (or, possibly, copies for primitive types) and packed tags on the stack and calling a formatting function. See Appendix B: Binary code comparison for a specific example.

Exposing the type-erased API rather than making it an implementation detail and only providing variadic functions allows applying the same technique to the user code. For example, consider a simple logging function that writes a formatted error message to clog:

template<class... Args>
  void log_error(int code, string_view fmt, const Args&... args) {
    clog << "Error " << code << ": " << format(fmt, args...);
  }

The code for this function will be generated for every combination of argument types which can be undesirable. However, if we use the type-erased API, there will be only one instance of the logging function while the wrapper function can be trivially inlined:

void vlog_error(int code, string_view fmt, format_args args) {
  clog << "Error " << code << ": " << vformat(fmt, args);
}

template<class... Args>
  inline void log_error(int code, string_view fmt, const Args&... args) {
    vlog_error(code, fmt, make_format_args(args...));
  }

The current design allows users to easily switch between the two approaches.

Compile-time processing of format strings

It is possible to parse format strings at compile time with constexpr functions which has been demonstrated in the reference implementation [14] and in [18]. Unfortunately a satisfactory API cannot be provided using existing C++17 features. Ideally we would like to use a function call API similar to the one proposed in this paper:

template<class String, class... Args>
  string format(String fmt, const Args&... args);

where String is some type representing a compile-time format string, for example

struct MyString {
  static constexpr string_view value() { return string_view("{}", 2); }
};

However, requiring a user to create a new type either manually or via a macro for every format string is not acceptable. P0424R1 "Reconsidering literal operator templates for strings" [15] provides a solution to this problem based on user-defined literal operators. If this or other similar proposal for compile-time strings is accepted into the standard, it should be easy to provide additional formatting APIs that make use of this feature. Obviously runtime checks will still be needed in cases where the format string is not known at compile time, but as shown in Appendix A: Benchmarks even with runtime parsing performance can be on par with or better than that of existing methods.

Compile-time processing of format strings can work with a user-defined type T if the formatter<T>::parse function is constexpr.

Argument visitation

The argument visitation API can be used to implement dynamic format specifiers, for example, width and precision passed as additional formatting arguments as opposed to being encoded in the format string itself:

int width = 10;
int precision = 3;
auto s = format("{0:{1}.{2}f}", 12.345678, width, precision);
// s == "    12.346"

An example of a user-defined formatter with dynamic width:

struct Answer {};

template<>
  struct formatter<Answer> {
    int width_arg_index = 0;

    // Parses dynamic width in the format "{<digit>}".
    auto parse(parse_context& parse_ctx) {
      auto iter = parse_ctx.begin();
      auto get_char = [&]() { return iter != parse_ctx.end() ? *iter : 0; };

      if (get_char() != '{')
        return iter;
      ++iter;
      char c = get_char();
      if (!std::isdigit(c) || (++iter, get_char()) != '}')
        throw format_error("invalid format");
      width_arg_index = c - '0';
      return ++iter;
    }

    auto format(Answer, format_context& format_ctx) {
      auto arg = format_ctx.args().get(width_arg_index);
      int width = visit_format_arg([](auto value) -> int {
        if constexpr (!std::is_integral_v<decltype(value)>)
          throw format_error("width is not integral");
        else if (value < 0 || value > std::numeric_limits<int>::max())
          throw format_error("invalid width");
        else
          return value;
        }, arg);
      return format_to(format_ctx.out(), "{:{}}", 42, width);
    }
  };

std::string s = format("{0:{1}}", Answer(), 10);
// s == "        42"

In many cases formatting can be delegated to standard formatters which makes manual handling of dynamic specifiers unneccessary, but the latter is still important for more complex cases.

This API can also be used to implement a custom formatting engine, such as the one compatible with the printf syntax, to provide some of the benefits of the current proposal to the legacy code.

Impact on existing code

The proposed formatting API is defined in the new header <format> and should have no impact on existing code.

Proposed wording

Add a new section in 19 [utilities].

Formatting utilities [format]

Header <format> synopsis [format.syn]

namespace std {
  // [format.error], class format_error
  class format_error;

  // [format.formatter], formatter
  template<class charT> class basic_parse_context;
  using parse_context = basic_parse_context<char>;
  using wparse_context = basic_parse_context<wchar_t>;
  
  template<class OutputIterator, class charT> class basic_format_context;
  using format_context = basic_format_context<unspecified, char>;
  using wformat_context = basic_format_context<unspecified, wchar_t>;

  template<class T, class charT = char> struct formatter;
  
  // [format.arguments], arguments
  template<class Context> class basic_format_arg;

  template<class Visitor, class Context>
    see below visit_format_arg(Visitor&& vis, basic_format_arg<Context> arg);

  template<class Context, class... Args> using format_arg_store = unspecified;

  template<class Context> class basic_format_args;
  using format_args = basic_format_args<format_context>;
  using wformat_args = basic_format_args<wformat_context>;

  template<class OutputIterator, class charT>
    using format_args_t = basic_format_args<basic_format_context<OutputIterator, charT>>;

  template<class Context, class... Args>
    format_arg_store<Context, Args...> make_format_args(const Args&... args);
  template<class... Args>
    format_arg_store<format_context, Args...> make_format_args(const Args&... args);
  template<class... Args>
    format_arg_store<wformat_context, Args...> make_wformat_args(const Args&... args);

  // [format.functions], formatting functions
  template<class... Args>
    string format(string_view fmt, const Args&... args);
  template<class... Args>
    wstring format(wstring_view fmt, const Args&... args);

  string vformat(string_view fmt, format_args args);
  wstring vformat(wstring_view fmt, wformat_args args);

  template<class OutputIterator, class... Args>
    OutputIterator format_to(OutputIterator out, string_view fmt,
                             const Args&... args);
  template<class OutputIterator, class... Args>
    OutputIterator format_to(OutputIterator out, wstring_view fmt,
                             const Args&... args);

  template<class OutputIterator>
    OutputIterator vformat_to(OutputIterator out, string_view fmt,
                              format_args_t<OutputIterator, char> args);
  template<class OutputIterator>
    OutputIterator vformat_to(OutputIterator out, wstring_view fmt,
                              format_args_t<OutputIterator, wchar_t> args);

  template<class OutputIterator, class Size>
    struct format_to_n_result {
      OutputIterator out;
      Size size;
    };
  
  template<class OutputIterator, class Size, class... Args>
    format_to_n_result<OutputIterator, Size>
      format_to_n(OutputIterator out, Size n, string_view fmt,
                  const Args&... args);
  template<class OutputIterator, class Size, class... Args>
    format_to_n_result<OutputIterator, Size>
      format_to_n(OutputIterator out, Size n, wstring_view fmt,
                  const Args&... args);

  template<class... Args>
    size_t formatted_size(string_view fmt, const Args&... args);
  template<class... Args>
    size_t formatted_size(wstring_view fmt, const Args&... args);
}

Class format_error [format.error]

namespace std {
  class format_error : public runtime_error {
  public:
    explicit format_error(const string& what_arg);
    explicit format_error(const char* what_arg);
  };
}

The class format_error defines the type of objects thrown as exceptions to report errors from the formatting library.

format_error(const string& what_arg);

Effects: Constructs an object of class format_error.

Ensures: strcmp(what(), what_arg.c_str()) == 0.

format_error(const char* what_arg);

Effects: Constructs an object of class format_error.

Ensures: strcmp(what(), what_arg) == 0.

Arguments [format.arguments]

Class template basic_format_arg
namespace std {
  template<class Context>
  class basic_format_arg {
  public:
    class handle;

    basic_format_arg();

    explicit operator bool() const noexcept;

    bool is_arithmetic() const;
    bool is_integral() const;
  };
}

The class template basic_format_arg describes an object that can refer to a formatting function argument. It is parameterized on a formatting context type (see [format.formatter]) and can hold a value of one of the following types:

where charT is typename Context::char_type. The value can be accessed via the visitation interface defined in the next section. basic_format_arg must preserve argument values, but does not have to distinguish between integral types other than charT and bool, e.g. an argument of type short can be stored as int and observed as int via the visitation interface. Similarly, basic_format_arg does not have to distinguish between floating-point types.

basic_format_arg();

Effects: Constructs a basic_format_arg object that doesn't refer to an argument.

Ensures: !(*this).

explicit operator bool() const noexcept;

Returns: true if and only if *this refers to an argument.

bool is_arithmetic() const;

Returns: true if and only if *this refers to an argument of an arithmetic type.

bool is_integral() const;

Returns: true if and only if *this refers to an argument of an integral type.

namespace std {
  template<class Context>
  class basic_format_arg<Context>::handle {
  public:
    void format(Context& ctx) const;
  };
}

The class handle describes an object that can refer to a formatting function argument of a user-defined type.

void format(Context& ctx) const;

Effects: Constructs a local variable of type typename Context::template formatter_type<T> (named f for exposition purposes), where T is the type of the object obj referred to by this handle. Executes ctx.parse_context().advance_to(f.parse(ctx.parse_context())). Then calls ctx.advance_to(f.format(obj, ctx)).

Argument visitation
template<class Visitor, class Context>
  see below visit_format_arg(Visitor&& vis, basic_format_arg<Context> arg);

Requires: The expression in the Effects section shall be a valid expression of the same type, for all alternative value types of a formatting argument. Otherwise, the program is ill-formed.

Effects: Let value be the value stored in arg. Returns INVOKE(std::forward(vis), value);.

Remarks: The return type is the common type of all possible INVOKE expressions in the Effects section. Since exact value types are implementation-defined, visitors should use type traits to handle multiple types.

[ Example:

  auto uint_value = visit_format_arg([](auto value) -> unsigned {
    if constexpr (std::is_unsigned_v<decltype(value)>)
      return value;
    return 0;
  }, arg);
end example ]

Class template basic_format_args
namespace std {
  template<class Context>
  class basic_format_args {
  public:
    using size_type = size_t;

    basic_format_args() noexcept;

    template<class... Args>
      basic_format_args(const format_arg_store<Context, Args...>& store);

    basic_format_arg<Context> get(size_type i) const;
  };
}

An object of type basic_format_args provides access to formatting arguments. Copying a basic_format_args object does not copy the arguments.

basic_format_args() noexcept;

Effects: Constructs an empty basic_format_args object.

Ensures: !get(0).

template<class... Args>
  basic_format_args(const format_arg_store<Context, Args...>& store);

Effects: Constructs a basic_format_args object that provides access to the arguments captured in store.

basic_format_arg<Context> get(size_type i) const;

Returns: A basic_format_arg object that represents an argument at index i if i < the number of arguments. Otherwise, returns an empty basic_format_arg object.

Function template make_format_args
template<class Context, class... Args>
  format_arg_store<Context, Args...> make_format_args(const Args&... args);
template<class... Args>
  format_arg_store<format_context, Args...> make_format_args(const Args&... args);

Effects: The function returns a format_arg_store object that refers to formatting arguments args.

Function template make_wformat_args
template<class... Args>
  format_arg_store<wformat_context, Args...> make_wformat_args(const Args&... args);

Effects: The function returns a format_arg_store object that refers to formatting arguments args.

Formatting functions [format.functions]

template<class... Args>
  string format(string_view fmt, const Args&... args);

Returns: vformat(fmt, make_format_args(args...)).

template<class... Args>
  wstring format(wstring_view fmt, const Args&... args);

Returns: vformat(fmt, make_wformat_args(args...)).

string vformat(string_view fmt, format_args args);
wstring vformat(wstring_view fmt, wformat_args args);

Effects: The function returns a string object constructed from the format string argument fmt with each replacement field substituted with the character representation of the argument it refers to, formatted according to the specification given in the field.

Returns: The formatted string.

Throws: format_error if fmt is not a valid format string.

template<class OutputIterator, class... Args>
  OutputIterator format_to(OutputIterator out, string_view fmt,
                           const Args&... args);

Returns: vformat_to(out, fmt, make_format_args(args...)).

template<class OutputIterator, class... Args>
  OutputIterator format_to(OutputIterator out, wstring_view fmt,
                           const Args&... args);

Returns: vformat_to(out, fmt, make_wformat_args(args...)).

template<class OutputIterator, class... Args>
  OutputIterator vformat_to(OutputIterator out, string_view fmt,
                            format_args_t<OutputIterator, char> args);
template<class OutputIterator, class... Args>
  OutputIterator vformat_to(OutputIterator out, wstring_view fmt,
                            format_args_t<OutputIterator, wchar_t> args);

Effects: The function writes to the range [out, out + size), where size is the output size, the format string fmt with each replacement field substituted with the character representation of the argument it refers to, formatted according to the specification given in the field.

Returns: The end of the output range.

Throws: format_error if fmt is not a valid format string.

template<class OutputIterator, class Size, class... Args>
  format_to_n_result<OutputIterator, Size>
    format_to_n(OutputIterator out, Size n, string_view fmt,
                const Args&... args);
template<class OutputIterator, class Size, class... Args>
  format_to_n_result<OutputIterator, Size>
    format_to_n(OutputIterator out, Size n, wstring_view fmt,
                const Args&... args);

Effects: The function writes to the range [out, out + n) the format string fmt with each replacement field substituted with the character representation of the argument it refers to, formatted according to the specification given in the field.

Returns: format_to_n_result containing the end of the output range and the total (not truncated) output size.

Throws: format_error if fmt is not a valid format string.

template<class... Args>
  size_t formatted_size(string_view fmt, const Args&... args);
template<class... Args>
  size_t formatted_size(wstring_view fmt, const Args&... args);

Effects: The function returns the number of characters in the output of format(fmt, args...) without constructing the formatted string.

Returns: The number of characters in the output of format(fmt, args...).

Throws: format_error if fmt is not a valid format string.

The fmt string consists of zero or more replacement fields, escape sequences, and ordinary multibyte characters. All ordinary multibyte characters are copied unchanged to the output. An escape sequence is one of {{ or }}; it is replaced with { or } respectively in the output. The syntax of replacement fields is as follows:

replacement-field ::= '{' [arg-id] [':' format-spec] '}'
arg-id            ::= integer
integer           ::= digit [integer]
digit             ::= '0'...'9'

The arg-id field specifies the index of the argument in args whose value is to be formatted and inserted into the output instead of the replacement field. The optional format-spec field specifies a non-default format for the replacement value.

If the numeric arg-ids in a format string are 0, 1, 2, ... in sequence, they can all be omitted (not just some) and the numbers 0, 1, 2, ... will be automatically used in that order. Mixing automatic and manual indexing is not allowed.

[ Example:

  string s0 = format("From {} to {}", "a", "b");   // OK: automatic indexing
  string s1 = format("From {1} to {0}", "a", "b"); // OK: manual indexing
  string s2 = format("From {0} to {}", "a", "b");  // Error: mixing automatic and manual indexing
  string s3 = format("From {} to {1}", "a", "b");  // Error: mixing automatic and manual indexing
end example ]

The format-spec field contains format specifications that define how the value should be presented, including such details as field width, alignment, padding, and decimal precision. Each type can define its own formatting mini-language or interpretation of the format-spec field. The syntax of format specifications is as follows:

format-spec     ::= std-format-spec | custom-format-spec
std-format-spec ::= [[fill] align] [sign] ['#'] ['0'] [width] ['.' precision] [type]
fill            ::= <a character other than '{' or '}'>
align           ::= '<' | '>' | '=' | '^'
sign            ::= '+' | '-' | ' '
width           ::= integer | '{' arg-id '}'
precision       ::= integer | '{' arg-id '}'
type            ::= 'a' | 'A' | 'b' | 'B' | 'c' | 'd' | 'e' | 'E' | 'f' | 'F' |
                    'g' | 'G' | 'n' | 'o' | 'p' | 's' | 'x' | 'X'

where std-format-spec defines a common formatting mini-language supported by built-in types and strings, while custom-format-spec is a placeholder for user-defined mini-languages. Some of the formatting options are only supported by arithmetic types.

The fill character can be any character other than '{' or '}'. The presence of a fill character is signaled by the character following it, which must be one of the alignment options. If the second character of format-spec is not a valid alignment option, then it is assumed that both the fill character and the alignment option are absent.

The meaning of the various alignment options is as follows:

OptionMeaning
'<' Forces the field to be left-aligned within the available space. This is the default for non-arithmetic types.
'>' Forces the field to be right-aligned within the available space. This is the default for arithmetic types.
'=' Forces the padding to be placed after the sign (if any) but before the digits. This is used for printing fields in the form +000000120. This alignment option is only valid for arithmetic types.
'^' Forces the field to be centered within the available space.

Note that unless a minimum field width is defined, the field width will be determined by the width of the content, meaning that the alignment option has no effect.

The sign option is only valid for arithmetic types, and can be one of the following:

OptionMeaning
'+' Indicates that a sign should be used for both positive as well as negative numbers.
'-' Indicates that a sign should be used only for negative numbers (this is the default behavior).
space Indicates that a leading space should be used for positive numbers, and a minus sign for negative numbers.

The '#' option causes the alternate form to be used for the conversion. This option is only valid for integer and floating-point types. For integers, when binary, octal, or hexadecimal output is used, this option adds the respective prefix "0b" ("0B"), "0", or "0x" ("0X") to the output value. Whether the prefix is lower-case or upper-case is determined by the case of the type format specifier. For floating-point numbers the alternate form causes the result of the conversion to always contain a decimal-point character, even if no digits follow it. Normally, a decimal-point character appears in the result of these conversions only if a digit follows it. In addition, for 'g' and 'G' conversions, trailing zeros are not removed from the result.

width is a decimal integer defining the minimum field width. If not specified, then the field width will be determined by the content.

Preceding the width field by a zero ('0') character enables sign-aware zero-padding for arithmetic types. This is equivalent to a fill character of '0' with an alignment type of '='.

The precision is a decimal number indicating how many digits should be displayed after the decimal point for a floating-point value formatted with 'f' and 'F', or before and after the decimal point for a floating-point value formatted with 'g' or 'G'. For non-arithmetic types the field indicates the maximum field size - in other words, how many characters will be used from the field content. The precision is not allowed for integer, character, boolean, and pointer values.

Finally, the type determines how the data should be presented.

The available string presentation types are:

TypeMeaning
's' String format. This is the default type for strings and may be omitted.
none The same as 's'.

The available character presentation types are:

TypeMeaning
'c' Character format. This is the default type for characters and may be omitted.
none The same as 'c'.

The available integer presentation types are:

TypeMeaning
'b' Binary format. Outputs the number in base 2. Using the '#' option with this type adds the prefix "0b" to the output value.
'B' Binary format. Outputs the number in base 2. Using the '#' option with this type adds the prefix "0B" to the output value.
'd' Decimal integer. Outputs the number in base 10.
'o' Octal format. Outputs the number in base 8. Using the '#' option with this type adds the prefix "0" to the output value.
'x' Hex format. Outputs the number in base 16, using lower-case letters for the digits above 9. Using the '#' option with this type adds the prefix "0x" to the output value.
'X' Hex format. Outputs the number in base 16, using upper-case letters for the digits above 9. Using the '#' option with this type adds the prefix "0X" to the output value.
'n' Number. This is the same as 'd' except that it uses the current locale to insert the appropriate number separator characters.
none The same as 'd'.

Integer presentation types can also be used with character and boolean values. Boolean values are formatted using textual representation, either true or false, if the presentation type is not specified.

The available presentation types for floating-point values are:

TypeMeaning
'a' Hexadecimal floating point format. Prints the number in base 16 with prefix "0x" and lower-case letters for digits above 9. Uses 'p' to indicate the exponent.
'A' Same as 'a' except it uses upper-case letters for the prefix, digits above 9 and to indicate the exponent.
'e' Exponent notation. Prints the number in scientific notation using the letter 'e' to indicate the exponent.
'E' Exponent notation. Same as 'e' except it uses an upper-case 'E' as the separator character.
'f' Fixed point. Displays the number as a fixed-point number.
'F' Fixed point. Same as 'f', but converts nan to NAN and inf to INF.
'g' General format. For a given precision p >= 1, this rounds the number to p significant digits and then formats the result in either fixed-point format or in scientific notation, depending on its magnitude. A precision of 0 is treated as equivalent to a precision of 1.
'G' General format. Same as 'g' except switches to 'E' if the number gets too large. The representations of infinity and NaN are uppercased, too.
'n' Number. This is the same as 'g', except that it uses the locale to insert the appropriate number separator characters.
none The same as 'g'.

The available presentation types for pointers are:

TypeMeaning
'p' Pointer format. This is the default type for pointers and may be omitted.
none The same as 'p'.

Formatter [format.formatter]

Class template basic_parse_context
namespace std {
  template<class charT>
  class basic_parse_context {
  public:
    using char_type = charT;
    using const_iterator = typename basic_string_view<charT>::const_iterator;
    using iterator = const_iterator;

    explicit constexpr basic_parse_context(basic_string_view<charT> fmt);

    constexpr const_iterator begin() const noexcept;
    constexpr const_iterator end() const noexcept;
    constexpr void advance_to(const_iterator it);

    constexpr size_t next_arg_id();
    constexpr void check_arg_id(size_t id);
  };
}

The class template basic_parse_context provides access the format string range being parsed and the argument counter for automatic indexing.

explicit constexpr basic_parse_context(basic_string_view<charT> fmt);

Effects: Constructs a parsing context from the format string fmt.

Ensures: begin() == fmt.begin(), end() == fmt.end().

constexpr const_iterator begin() const noexcept;

Returns: An iterator referring to the first character in the format string range being parsed.

constexpr const_iterator end() const noexcept;

Returns: An iterator referring to the position one past the last character in the format string range being parsed.

void advance_to(iterator it);

Effects: Advances the beginning of the parsed format string range to it.

Requires: end() shall be reachable from it.

Ensures: begin() == it.

constexpr size_t next_arg_id();

Returns: The next argument identifier starting from 0. Each subsequent call to next_arg_id gives an identifier which is 1 greater than the one returned from the previous call for the same context object.

Throws: format_error if check_arg_id has been called earlier which indicates mixing of automatic and manual argument indexing.

constexpr void check_arg_id(size_t id);

Throws: format_error if next_arg_id has been called earlier which indicates mixing of automatic and manual argument indexing.

Class template basic_format_context
namespace std {
  template<class OutputIterator, class charT>
  class basic_format_context {
  public:
    using iterator = OutputIterator;
    using char_type = charT;
    using args_type = basic_format_args<basic_format_context>;

    template<class T>
      using formatter_type = formatter<T>;

    basic_parse_context<charT>& parse_context();

    iterator out();
    void advance_to(iterator it);

    args_type args() const;
  };
}

The class template basic_format_context provides access to format string parsing context, output iterator and formatting arguments.

basic_parse_context<charT>& parse_context();

Returns: A reference to the format string parsing context.

iterator out();

Returns: An iterator referring to the current position in the output range.

void advance_to(iterator it);

Effects: Advances the current position in the output range to it.

Ensures: out() == it.

args_type args() const;

Returns: A copy of the args_type object that provides access to formatting arguments.

Class template formatter

The functions defined in [format.functions] use specializations of the class template formatter to format individual arguments.

Each specialization of formatter is either enabled or disabled, as described below. [ Note: Enabled specializations meet the Formatter requirements, and disabled specializations do not. — end note ] Each header that declares the template formatter provides enabled specializations formatter<const char*, char>, formatter<const wchar_t*, wchar_t> as well as specialiazations of formatter for nullptr_t, const void*, and all cv-unqualified arithmetic types. For any type T for which neither the library nor the user provides an explicit or partial specialization of the class template formatter, formatter<T> is disabled.

If the library provides an explicit or partial specialization of formatter<T>, that specialization is enabled except as noted otherwise.

If F is a disabled specialization of formatter, these values are false: is_default_constructible_v<F>, is_copy_constructible_v<F>, is_move_constructible_v<F>, is_copy_assignable_v<F>, is_move_assignable_v<F>.

An enabled specialization formatter<T> will satisfy the Formatter requirements (Table 1), with T as the formatted argument type, the DefaultConstructible requirements, and the CopyAssignable requirements.

Formatter requirements

A type F meets the Formatter requirements if:

Given character type charT, output iterator type OutputIterator, and formatted argument type T, in Table 1 f is a value of type F, u is an lvalue of type T, t is a value of a type convertible to (possibly const) T, pc is an lvalue of type basic_parse_context<charT> (denoted by PC), and fc is an lvalue of type basic_format_context<OutputIterator, charT> (denoted by FC). pc.begin() points to the beginning of the format-spec ([format.syntax]) portion of the format string. If format-spec is empty then either pc.begin() == pc.end() or *pc.begin() == '}'.

Table 1 — Formatter requirements
ExpressionReturn typeRequirement
f.parse(pc) PC::iterator Shall parse format-spec for type T, store the parsed specifiers in *this, and return an iterator past the end of the parsed range.
f.format(t, fc) FC::iterator Shall format t according to the specifiers stored in *this, write the output to fc.out() and return an iterator past the end of the output range.
f.format(u, fc) FC::iterator Shall not modify u.

Related work

The Boost Format library [8] is an established formatting library that uses printf-like format string syntax with extensions. The main differences between this library and the current proposal are:

A printf-like Interface for the Streams Library [10] is similar to the Boost Format library but uses variadic templates instead of operator%. Unfortunately it hasn't been updated since 2013 and the same arguments about format string syntax apply to it.

The FastFormat library [11] is another well-known formatting library. Similarly to this proposal, FastFormat uses brace-delimited format specifiers, but otherwise the format string syntax is different and the library has significant limitations [12]:

Three features that have no hope of being accommodated within the current design are:
  • Leading zeros (or any other non-space padding)
  • Octal/hexadecimal encoding
  • Runtime width/alignment specification

Formatting facilities of the Folly library [13] are the closest to the current proposal. Folly also uses Python-like format string syntax nearly identical to the one described here. However, the API details are quite different. The current proposal tries to address performance and code bloat issues that are largely ignored by Folly Format. For instance formatting functions in Folly Format are parameterized on all argument types while in this proposal, only the inlined wrapper functions are, which results in much smaller binary code and better compile times.

Implementation

An implementation of this proposal is available in the std branch of the open-source fmt library [14].

Acknowledgements

Thanks to Beman Dawes, Bengt Gustafsson, Eric Niebler, Jason McKesson, Jeffrey Yasskin, Joël Lamotte, Lee Howes, Louis Dionne, Matt Clarke, Michael Park, Sergey Ignatchenko, Thiago Macieira, Zach Laine, Zhihao Yuan and participants of the Library Evolution Working Group for their feedback, support, constructive criticism and contributions to the proposal. Special thanks to Howard Hinnant who encouraged me to write the proposal and gave useful early advice on how to go about it.

The format string syntax is based on the Python documentation [3].

References

[1] The fprintf function. ISO/IEC 9899:2011. 7.21.6.1.
[2] fprintf, printf, snprintf, sprintf - print formatted output. The Open Group Base Specifications Issue 6 IEEE Std 1003.1, 2004 Edition.
[3] 6.1.3. Format String Syntax. Python 3.5.2 documentation.
[4] String.Format Method. .NET Framework Class Library.
[5] Module std::fmt. The Rust Standard Library.
[6] Format Specification Syntax: printf and wprintf Functions. C++ Language and Standard Libraries.
[7] 10.4.2 Rearranging printf Arguments. The GNU Awk User's Guide.
[8] Boost Format library. Boost 1.63 documentation.
[9] Speed Test. The fmt library repository.
[10] A printf-like Interface for the Streams Library (revision 1).
[11] The FastFormat library website.
[12] An Introduction to Fast Format (Part 1): The State of the Art. Overload Journal #89 - February 2009
[13] The folly library repository.
[14] The fmt library repository.
[15] P0424: Reconsidering literal operator templates for strings.
[16] D0355: Extending <chrono> to Calendars and Time Zones.
[17] P0067: Elementary string conversions.
[18] MPark.Format: Compile-time Checked, Type-Safe Formatting in C++14.
[19] Google Benchmark: A microbenchmark support library.

Appendix A: Benchmarks

To demonstrate that the formatting functions described in this paper can be implemented efficiently, we compare the reference implementation [14] of format and format_to to sprintf, ostringstream and to_string on the following benchmark. This benchmark generates a set of integers with random numbers of digits, applies each method to convert each integer into a string (either std::string or a char buffer depending on the API) and uses the Google Benchmark library [19] to measure timings:
  #include <algorithm>
  #include <cmath>
  #include <cstdio>
  #include <limits>
  #include <sstream>
  #include <string>
  #include <utility>
  #include <vector>

  #include <benchmark/benchmark.h>
  #include <fmt/format.h>

  // Returns a pair with the smallest and the largest value of integral type T
  // with the given number of digits.
  template<typename T>
  std::pair<T, T> range(int num_digits) {
    T first = std::pow(T(10), num_digits - 1);
    int max_digits = std::numeric_limits<T>::digits10 + 1;
    T last = num_digits < max_digits ? first * 10 - 1 :
                                 std::numeric_limits<T>::max();
    return {num_digits > 1 ? first : 0, last};
  }

  // Generates values of integral type T with random number of digits.
  template<typename T>
  std::vector<T> generate_random_data(int numbers_per_digit) {
    int max_digits = std::numeric_limits<T>::digits10 + 1;
    std::vector<T> data;
    data.reserve(max_digits * numbers_per_digit);
    for (int i = 1; i <= max_digits; ++i) {
      auto r = range<T>(i);
      auto value = r.first;
      std::generate_n(std::back_inserter(data), numbers_per_digit, [=]() mutable {
        T result = value;
        value = value < r.second ? value + 1 : r.first;
        return result;
      });
    }
    std::random_shuffle(data.begin(), data.end());
    return data;
  }

  auto data = generate_random_data<int>(1000);

  void sprintf(benchmark::State &s) {
    size_t result = 0;
    while (s.KeepRunning()) {
      for (auto i: data) {
        char buffer[12];
        result += std::sprintf(buffer, "%d", i);
      }
    }
    benchmark::DoNotOptimize(result);
  }
  BENCHMARK(sprintf);

  void ostringstream(benchmark::State &s) {
    size_t result = 0;
    while (s.KeepRunning()) {
      for (auto i: data) {
        std::ostringstream ss;
        ss << i;
        result += ss.str().size();
      }
    }
    benchmark::DoNotOptimize(result);
  }
  BENCHMARK(ostringstream);

  void to_string(benchmark::State &s) {
    size_t result = 0;
    while (s.KeepRunning()) {
      for (auto i: data)
        result += std::to_string(i).size();
    }
    benchmark::DoNotOptimize(result);
  }
  BENCHMARK(to_string);

  void format(benchmark::State &s) {
    size_t result = 0;
    while (s.KeepRunning()) {
      for (auto i: data)
        result += fmt::format("{}", i).size();
    }
    benchmark::DoNotOptimize(result);
  }
  BENCHMARK(format);

  void format_to(benchmark::State &s) {
    size_t result = 0;
    while (s.KeepRunning()) {
      for (auto i: data) {
        char buffer[12];
        result += fmt::format_to(buffer, "{}", i) - buffer;
      }
    }
    benchmark::DoNotOptimize(result);
  }
  BENCHMARK(format_to);

  BENCHMARK_MAIN();
The benchmark was compiled with clang (Apple LLVM version 9.0.0 clang-900.0.39.2) with -O3 -DNDEBUG and run on a macOS system. Below are the results:
  Run on (4 X 3100 MHz CPU s)
  2018-01-27 07:12:00
  Benchmark              Time           CPU Iterations
  ----------------------------------------------------
  sprintf           882311 ns     881076 ns        781
  ostringstream    2892035 ns    2888975 ns        242
  to_string        1167422 ns    1166831 ns        610
  format            675636 ns     674382 ns       1045
  format_to         499376 ns     498996 ns       1263
The format and format_to functions show much better performance than the other methods. The format function that constructs std::string is even 30% faster than the system's version of sprintf that uses stack-allocated char buffer. format_to with a stack-allocated buffer is ~60% faster than sprintf.

Appendix B: Binary code comparison

In this section we compare per-call binary code size between the reference implementation that uses techniques described in section Binary footprint and standard formatting facilities. All the code snippets are compiled with clang (Apple LLVM version 9.0.0 clang-900.0.39.2) with -O3 -DNDEBUG -c -std=c++14 and the resulted binaries are disassembled with objdump -S:
  void consume(const char*);

  void sprintf_test() {
    char buffer[100];
    sprintf(buffer, "The answer is %d.", 42);
    consume(buffer);
  }

  __Z12sprintf_testv:
         0:       55      pushq   %rbp
         1:       48 89 e5        movq    %rsp, %rbp
         4:       53      pushq   %rbx
         5:       48 83 ec 78     subq    $120, %rsp
         9:       48 8b 05 00 00 00 00    movq    (%rip), %rax
        10:       48 8b 00        movq    (%rax), %rax
        13:       48 89 45 f0     movq    %rax, -16(%rbp)
        17:       48 8d 35 37 00 00 00    leaq    55(%rip), %rsi
        1e:       48 8d 5d 80     leaq    -128(%rbp), %rbx
        22:       ba 2a 00 00 00  movl    $42, %edx
        27:       31 c0   xorl    %eax, %eax
        29:       48 89 df        movq    %rbx, %rdi
        2c:       e8 00 00 00 00  callq   0 <__Z12sprintf_testv+0x31>
        31:       48 89 df        movq    %rbx, %rdi
        34:       e8 00 00 00 00  callq   0 <__Z12sprintf_testv+0x39>
        39:       48 8b 05 00 00 00 00    movq    (%rip), %rax
        40:       48 8b 00        movq    (%rax), %rax
        43:       48 3b 45 f0     cmpq    -16(%rbp), %rax
        47:       75 07   jne     7 <__Z12sprintf_testv+0x50>
        49:       48 83 c4 78     addq    $120, %rsp
        4d:       5b      popq    %rbx
        4e:       5d      popq    %rbp
        4f:       c3      retq
        50:       e8 00 00 00 00  callq   0 <__Z12sprintf_testv+0x55>

  void format_test() {
    consume(format("The answer is {}.", 42).c_str());
  }

  __Z11format_testv:
         0:       55      pushq   %rbp
         1:       48 89 e5        movq    %rsp, %rbp
         4:       53      pushq   %rbx
         5:       48 83 ec 28     subq    $40, %rsp
         9:       48 c7 45 d0 2a 00 00 00         movq    $42, -48(%rbp)
        11:       48 8d 35 f4 83 01 00    leaq    99316(%rip), %rsi
        18:       48 8d 7d e0     leaq    -32(%rbp), %rdi
        1c:       4c 8d 45 d0     leaq    -48(%rbp), %r8
        20:       ba 11 00 00 00  movl    $17, %edx
        25:       b9 02 00 00 00  movl    $2, %ecx
        2a:       e8 00 00 00 00  callq   0 <__Z11format_testv+0x2F>
        2f:       f6 45 e0 01     testb   $1, -32(%rbp)
        33:       48 8d 7d e1     leaq    -31(%rbp), %rdi
        37:       48 0f 45 7d f0  cmovneq -16(%rbp), %rdi
        3c:       e8 00 00 00 00  callq   0 <__Z11format_testv+0x41>
        41:       f6 45 e0 01     testb   $1, -32(%rbp)
        45:       74 09   je      9 <__Z11format_testv+0x50>
        47:       48 8b 7d f0     movq    -16(%rbp), %rdi
        4b:       e8 00 00 00 00  callq   0 <__Z11format_testv+0x50>
        50:       48 83 c4 28     addq    $40, %rsp
        54:       5b      popq    %rbx
        55:       5d      popq    %rbp
        56:       c3      retq
        57:       48 89 c3        movq    %rax, %rbx
        5a:       f6 45 e0 01     testb   $1, -32(%rbp)
        5e:       74 09   je      9 <__Z11format_testv+0x69>
        60:       48 8b 7d f0     movq    -16(%rbp), %rdi
        64:       e8 00 00 00 00  callq   0 <__Z11format_testv+0x69>
        69:       48 89 df        movq    %rbx, %rdi
        6c:       e8 00 00 00 00  callq   0 <__Z11format_testv+0x71>
        71:       66 66 66 66 66 66 2e 0f 1f 84 00 00 00 00 00    nopw    %cs:(%rax,%rax)

  void ostringstream_test() {
    std::ostringstream ss;
    ss << "The answer is " << 42 << ".";
    consume(ss.str().c_str());
  }

  __Z18ostringstream_testv:
         0:	55 	pushq	%rbp
         1:	48 89 e5 	movq	%rsp, %rbp
         4:	41 57 	pushq	%r15
         6:	41 56 	pushq	%r14
         8:	41 55 	pushq	%r13
         a:	41 54 	pushq	%r12
         c:	53 	pushq	%rbx
         d:	48 81 ec 38 01 00 00 	subq	$312, %rsp
        14:	4c 8d b5 18 ff ff ff 	leaq	-232(%rbp), %r14
        1b:	4c 8d a5 b0 fe ff ff 	leaq	-336(%rbp), %r12
        22:	48 8b 05 00 00 00 00 	movq	(%rip), %rax
        29:	48 8d 48 18 	leaq	24(%rax), %rcx
        2d:	48 89 8d a8 fe ff ff 	movq	%rcx, -344(%rbp)
        34:	48 83 c0 40 	addq	$64, %rax
        38:	48 89 85 18 ff ff ff 	movq	%rax, -232(%rbp)
        3f:	4c 89 f7 	movq	%r14, %rdi
        42:	4c 89 e6 	movq	%r12, %rsi
        45:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x4A>
        4a:	48 c7 45 a0 00 00 00 00 	movq	$0, -96(%rbp)
        52:	c7 45 a8 ff ff ff ff 	movl	$4294967295, -88(%rbp)
        59:	48 8b 1d 00 00 00 00 	movq	(%rip), %rbx
        60:	4c 8d 6b 18 	leaq	24(%rbx), %r13
        64:	4c 89 ad a8 fe ff ff 	movq	%r13, -344(%rbp)
        6b:	48 83 c3 40 	addq	$64, %rbx
        6f:	48 89 9d 18 ff ff ff 	movq	%rbx, -232(%rbp)
        76:	4c 89 e7 	movq	%r12, %rdi
        79:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x7E>
        7e:	4c 8b 3d 00 00 00 00 	movq	(%rip), %r15
        85:	49 83 c7 10 	addq	$16, %r15
        89:	4c 89 bd b0 fe ff ff 	movq	%r15, -336(%rbp)
        90:	48 c7 85 08 ff ff ff 00 00 00 00 	movq	$0, -248(%rbp)
        9b:	48 c7 85 00 ff ff ff 00 00 00 00 	movq	$0, -256(%rbp)
        a6:	48 c7 85 f8 fe ff ff 00 00 00 00 	movq	$0, -264(%rbp)
        b1:	48 c7 85 f0 fe ff ff 00 00 00 00 	movq	$0, -272(%rbp)
        bc:	c7 85 10 ff ff ff 10 00 00 00 	movl	$16, -240(%rbp)
        c6:	0f 57 c0 	xorps	%xmm0, %xmm0
        c9:	0f 29 45 b0 	movaps	%xmm0, -80(%rbp)
        cd:	48 c7 45 c0 00 00 00 00 	movq	$0, -64(%rbp)
        d5:	48 8d 75 b0 	leaq	-80(%rbp), %rsi
        d9:	4c 89 e7 	movq	%r12, %rdi
        dc:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0xE1>
        e1:	f6 45 b0 01 	testb	$1, -80(%rbp)
        e5:	74 09 	je	9 <__Z18ostringstream_testv+0xF0>
        e7:	48 8b 7d c0 	movq	-64(%rbp), %rdi
        eb:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0xF0>
        f0:	48 8d 35 dd 10 00 00 	leaq	4317(%rip), %rsi
        f7:	48 8d bd a8 fe ff ff 	leaq	-344(%rbp), %rdi
        fe:	ba 0e 00 00 00 	movl	$14, %edx
       103:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x108>
       108:	be 2a 00 00 00 	movl	$42, %esi
       10d:	48 89 c7 	movq	%rax, %rdi
       110:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x115>
       115:	48 8d 35 c7 10 00 00 	leaq	4295(%rip), %rsi
       11c:	ba 01 00 00 00 	movl	$1, %edx
       121:	48 89 c7 	movq	%rax, %rdi
       124:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x129>
       129:	48 8d 7d b0 	leaq	-80(%rbp), %rdi
       12d:	4c 89 e6 	movq	%r12, %rsi
       130:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x135>
       135:	f6 45 b0 01 	testb	$1, -80(%rbp)
       139:	48 8d 7d b1 	leaq	-79(%rbp), %rdi
       13d:	48 0f 45 7d c0 	cmovneq	-64(%rbp), %rdi
       142:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x147>
       147:	f6 45 b0 01 	testb	$1, -80(%rbp)
       14b:	74 09 	je	9 <__Z18ostringstream_testv+0x156>
       14d:	48 8b 7d c0 	movq	-64(%rbp), %rdi
       151:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x156>
       156:	4c 89 ad a8 fe ff ff 	movq	%r13, -344(%rbp)
       15d:	48 89 9d 18 ff ff ff 	movq	%rbx, -232(%rbp)
       164:	4c 89 bd b0 fe ff ff 	movq	%r15, -336(%rbp)
       16b:	f6 85 f0 fe ff ff 01 	testb	$1, -272(%rbp)
       172:	74 0c 	je	12 <__Z18ostringstream_testv+0x180>
       174:	48 8b bd 00 ff ff ff 	movq	-256(%rbp), %rdi
       17b:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x180>
       180:	4c 89 e7 	movq	%r12, %rdi
       183:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x188>
       188:	48 8b 35 00 00 00 00 	movq	(%rip), %rsi
       18f:	48 83 c6 08 	addq	$8, %rsi
       193:	48 8d bd a8 fe ff ff 	leaq	-344(%rbp), %rdi
       19a:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x19F>
       19f:	4c 89 f7 	movq	%r14, %rdi
       1a2:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x1A7>
       1a7:	48 81 c4 38 01 00 00 	addq	$312, %rsp
       1ae:	5b 	popq	%rbx
       1af:	41 5c 	popq	%r12
       1b1:	41 5d 	popq	%r13
       1b3:	41 5e 	popq	%r14
       1b5:	41 5f 	popq	%r15
       1b7:	5d 	popq	%rbp
       1b8:	c3 	retq
       1b9:	48 89 45 d0 	movq	%rax, -48(%rbp)
       1bd:	f6 45 b0 01 	testb	$1, -80(%rbp)
       1c1:	74 3b 	je	59 <__Z18ostringstream_testv+0x1FE>
       1c3:	48 8b 7d c0 	movq	-64(%rbp), %rdi
       1c7:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x1CC>
       1cc:	eb 30 	jmp	48 <__Z18ostringstream_testv+0x1FE>
       1ce:	eb 2a 	jmp	42 <__Z18ostringstream_testv+0x1FA>
       1d0:	48 89 45 d0 	movq	%rax, -48(%rbp)
       1d4:	f6 45 b0 01 	testb	$1, -80(%rbp)
       1d8:	74 39 	je	57 <__Z18ostringstream_testv+0x213>
       1da:	48 8b 7d c0 	movq	-64(%rbp), %rdi
       1de:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x1E3>
       1e3:	f6 85 f0 fe ff ff 01 	testb	$1, -272(%rbp)
       1ea:	75 30 	jne	48 <__Z18ostringstream_testv+0x21C>
       1ec:	eb 3a 	jmp	58 <__Z18ostringstream_testv+0x228>
       1ee:	48 89 45 d0 	movq	%rax, -48(%rbp)
       1f2:	eb 3c 	jmp	60 <__Z18ostringstream_testv+0x230>
       1f4:	48 89 45 d0 	movq	%rax, -48(%rbp)
       1f8:	eb 4d 	jmp	77 <__Z18ostringstream_testv+0x247>
       1fa:	48 89 45 d0 	movq	%rax, -48(%rbp)
       1fe:	4c 89 ad a8 fe ff ff 	movq	%r13, -344(%rbp)
       205:	48 89 9d 18 ff ff ff 	movq	%rbx, -232(%rbp)
       20c:	4c 89 bd b0 fe ff ff 	movq	%r15, -336(%rbp)
       213:	f6 85 f0 fe ff ff 01 	testb	$1, -272(%rbp)
       21a:	74 0c 	je	12 <__Z18ostringstream_testv+0x228>
       21c:	48 8b bd 00 ff ff ff 	movq	-256(%rbp), %rdi
       223:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x228>
       228:	4c 89 e7 	movq	%r12, %rdi
       22b:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x230>
       230:	48 8b 35 00 00 00 00 	movq	(%rip), %rsi
       237:	48 83 c6 08 	addq	$8, %rsi
       23b:	48 8d bd a8 fe ff ff 	leaq	-344(%rbp), %rdi
       242:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x247>
       247:	4c 89 f7 	movq	%r14, %rdi
       24a:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x24F>
       24f:	48 8b 7d d0 	movq	-48(%rbp), %rdi
       253:	e8 00 00 00 00 	callq	0 <__Z18ostringstream_testv+0x258>
       258:	0f 1f 84 00 00 00 00 00 	nopl	(%rax,%rax)

The code generated for the format_test function that uses the reference implementation of the format function described in this proposal is several times smaller than the ostringstream code and only 40% larger than the one generated for sprintf which is a moderate price to pay for full type and memory safety.

The following factors contribute to the difference in binary code size between format and sprintf:

We can exclude the first two factors from the experiment by mimicking parts of the sprintf API:
int vraw_format(char* buffer, const char* format, format_args args);

template<typename... Args>
inline int raw_format(char* buffer, const char* format, const Args&... args) {
  return vraw_format(buffer, format, make_format_args(args...));
}

void raw_format_test() {
  char buffer[100];
  raw_format(buffer, "The answer is {}.", 42);
}

__Z15raw_format_testv:
       0:       55      pushq   %rbp
       1:       48 89 e5        movq    %rsp, %rbp
       4:       48 81 ec 80 00 00 00    subq    $128, %rsp
       b:       48 8b 05 00 00 00 00    movq    (%rip), %rax
      12:       48 8b 00        movq    (%rax), %rax
      15:       48 89 45 f8     movq    %rax, -8(%rbp)
      19:       48 c7 45 80 2a 00 00 00         movq    $42, -128(%rbp)
      21:       48 8d 35 24 12 00 00    leaq    4644(%rip), %rsi
      28:       48 8d 7d 90     leaq    -112(%rbp), %rdi
      2c:       48 8d 4d 80     leaq    -128(%rbp), %rcx
      30:       ba 02 00 00 00  movl    $2, %edx
      35:       e8 00 00 00 00  callq   0 <__Z15raw_format_testv+0x3A>
      3a:       48 8b 05 00 00 00 00    movq    (%rip), %rax
      41:       48 8b 00        movq    (%rax), %rax
      44:       48 3b 45 f8     cmpq    -8(%rbp), %rax
      48:       75 09   jne     9 <__Z15raw_format_testv+0x53>
      4a:       48 81 c4 80 00 00 00    addq    $128, %rsp
      51:       5d      popq    %rbp
      52:       c3      retq
      53:       e8 00 00 00 00  callq   0 <__Z15raw_format_testv+0x58>
      58:       0f 1f 84 00 00 00 00 00         nopl    (%rax,%rax)
This shows that passing formatting arguments adds very little overhead and is comparable with sprintf.