{fmt} is an open-source formatting library for C++. It can be used as a safe alternative to printf or as a fast alternative to IOStreams.
This is a development branch, released versions are available from the Releases page.
- Two APIs: faster concatenation-based write API and slower, but still very fast, replacement-based format API with positional arguments for localization.
- Write API similar to the one used by IOStreams but stateless allowing faster implementation.
- Format API with format string syntax similar to the one used by str.format in Python.
- Safe printf implementation including the POSIX extension for positional arguments.
- Support for user-defined types.
- High speed: performance of the format API is close to that of glibc's printf and better than the performance of IOStreams. See Speed tests and Fast integer to string conversion in C++.
- Small code size both in terms of source code (the core library consists of two header files and a single source file) and compiled code. See Compile time and code bloat.
- Reliability: the library has an extensive set of unit tests.
- Safety: the library is fully type safe, errors in format strings can be reported at compile time, automatic memory management prevents buffer overflow errors.
- Ease of use: small self-contained code base, no external dependencies, permissive BSD license
- Portability with consistent output across platforms and support for older compilers.
- Clean warning-free codebase even on high warning levels
(
-Wall -Wextra -pedantic). - Support for wide strings.
- Optional header-only configuration enabled with the
FMT_HEADER_ONLYmacro.
See the documentation for more details.
This prints Hello, world! to stdout:
fmt::print("Hello, {}!", "world"); // uses Python-like format string syntax
fmt::printf("Hello, %s!", "world"); // uses printf format string syntaxArguments can be accessed by position and arguments' indices can be repeated:
std::string s = fmt::format("{0}{1}{0}", "abra", "cad");
// s == "abracadabra"Format strings can be checked at compile time:
// test.cc
using namespace fmt::literals;
std::string s = "{2}"_format(42);$ g++ -Iinclude test.cc -std=c++14
...
test.cc:5:31: note: in instantiation of function template specialization
'fmt::internal::udl_formatter<char, '{', '2', '}'>::operator()<int>' requested
here
std::string s = "{2}"_format(42);
^
include/fmt/format.h:3838:7: note: non-constexpr function 'on_error' cannot be
used in a constant expression
on_error("argument index out of range");
^
{fmt} can be used as a safe portable replacement for itoa:
fmt::MemoryWriter w;
w << 42; // replaces itoa(42, buffer, 10)
w << fmt::hex(42); // replaces itoa(42, buffer, 16)
// access the string using w.str() or w.c_str()An object of any user-defined type for which there is an overloaded
std::ostream insertion operator (operator<<) can be formatted:
#include "fmt/ostream.h"
class Date {
int year_, month_, day_;
public:
Date(int year, int month, int day) : year_(year), month_(month), day_(day) {}
friend std::ostream &operator<<(std::ostream &os, const Date &d) {
return os << d.year_ << '-' << d.month_ << '-' << d.day_;
}
};
std::string s = fmt::format("The date is {}", Date(2012, 12, 9));
// s == "The date is 2012-12-9"You can create your own functions similar to format and print which take arbitrary arguments:
// Prints formatted error message.
void vreport_error(const char *format, fmt::args args) {
fmt::print("Error: ");
fmt::vprint(format, args);
}
template <typename... Args>
void report_error(const char *format, const Args & ... args) {
vreport_error(format, fmt::make_args(args));
}
report_error("file not found: {}", path);Note that vreport_error is not parameterized on argument types which can
improve compile times and reduce code size compared to fully parameterized version.
- 0 A.D.: A free, open-source, cross-platform real-time strategy game
- AMPL/MP: An open-source library for mathematical programming
- CUAUV: Cornell University's autonomous underwater vehicle
- HarpyWar/pvpgn: Player vs Player Gaming Network with tweaks
- KBEngine: An open-source MMOG server engine
- Keypirinha: A semantic launcher for Windows
- Kodi (formerly xbmc): Home theater software
- Lifeline: A 2D game
- Drake: A planning, control, and analysis toolbox for nonlinear dynamical systems (MIT)
- Envoy: C++ L7 proxy and communication bus (Lyft)
- FiveM: a modification framework for GTA V
- MongoDB Smasher: A small tool to generate randomized datasets
- OpenSpace: An open-source astrovisualization framework
- PenUltima Online (POL): An MMO server, compatible with most Ultima Online clients
- quasardb: A distributed, high-performance, associative database
- readpe: Read Portable Executable
- redis-cerberus: A Redis cluster proxy
- Saddy: Small crossplatform 2D graphic engine
- Salesforce Analytics Cloud: Business intelligence software
- Scylla: A Cassandra-compatible NoSQL data store that can handle 1 million transactions per second on a single server
- Seastar: An advanced, open-source C++ framework for high-performance server applications on modern hardware
- spdlog: Super fast C++ logging library
- Stellar: Financial platform
- Touch Surgery: Surgery simulator
- TrinityCore: Open-source MMORPG framework
If you are aware of other projects using this library, please let me know by email or by submitting an issue.
So why yet another formatting library?
There are plenty of methods for doing this task, from standard ones like the printf family of function and IOStreams to Boost Format library and FastFormat. The reason for creating a new library is that every existing solution that I found either had serious issues or didn't provide all the features I needed.
The good thing about printf is that it is pretty fast and readily available being a part of the C standard library. The main drawback is that it doesn't support user-defined types. Printf also has safety issues although they are mostly solved with __attribute__ ((format (printf, ...)) in GCC. There is a POSIX extension that adds positional arguments required for i18n to printf but it is not a part of C99 and may not be available on some platforms.
The main issue with IOStreams is best illustrated with an example:
std::cout << std::setprecision(2) << std::fixed << 1.23456 << "\n";which is a lot of typing compared to printf:
printf("%.2f\n", 1.23456);Matthew Wilson, the author of FastFormat, referred to this situation with IOStreams as "chevron hell". IOStreams doesn't support positional arguments by design.
The good part is that IOStreams supports user-defined types and is safe although error reporting is awkward.
This is a very powerful library which supports both printf-like format strings and positional arguments. The main its drawback is performance. According to various benchmarks it is much slower than other methods considered here. Boost Format also has excessive build times and severe code bloat issues (see Benchmarks).
This is an interesting library which is fast, safe and has positional arguments. However it has significant limitations, citing its author:
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
It is also quite big and has a heavy dependency, STLSoft, which might be too restrictive for using it in some projects.
SafeFormat is a formatting library which uses printf-like format strings
and is type safe. It doesn't support user-defined types or positional
arguments. It makes unconventional use of operator() for passing
format arguments.
This library supports printf-like format strings and is very small and fast. Unfortunately it doesn't support positional arguments and wrapping it in C++98 is somewhat difficult. Also its performance and code compactness are limited by IOStreams.
This is not really a formatting library but I decided to include it here
for completeness. As IOStreams it suffers from the problem of mixing
verbatim text with arguments. The library is pretty fast, but slower
on integer formatting than fmt::Writer on Karma's own benchmark,
see Fast integer to string conversion in C++.
The following speed tests results were generated by building
tinyformat_test.cpp on Ubuntu GNU/Linux 14.04.1 with
g++-4.8.2 -O3 -DSPEED_TEST -DHAVE_FORMAT, and taking the best of three
runs. In the test, the format string "%0.10f:%04d:%+g:%s:%p:%c:%%\n" or
equivalent is filled 2000000 times with output sent to /dev/null; for
further details see the source.
| Library | Method | Run Time, s |
|---|---|---|
| libc | printf | 1.35 |
| libc++ | std::ostream | 3.42 |
| fmt 534bff7 | fmt::print | 1.56 |
| tinyformat 2.0.1 | tfm::printf | 3.73 |
| Boost Format 1.54 | boost::format | 8.44 |
| Folly Format | folly::format | 2.54 |
As you can see boost::format is much slower than the alternative methods; this
is confirmed by other tests.
Tinyformat is quite good coming close to IOStreams. Unfortunately tinyformat
cannot be faster than the IOStreams because it uses them internally.
Performance of fmt is close to that of printf, being faster than printf on integer
formatting,
but slower on floating-point formatting which dominates this benchmark.
The script bloat-test.py
from format-benchmark
tests compile time and code bloat for nontrivial projects.
It generates 100 translation units and uses printf() or its alternative
five times in each to simulate a medium sized project. The resulting
executable size and compile time (Apple LLVM version 8.1.0 (clang-802.0.42),
macOS Sierra, best of three) is shown in the following tables.
Optimized build (-O3)
| Method | Compile Time, s | Executable size, KiB | Stripped size, KiB |
|---|---|---|---|
| printf | 2.7 | 29 | 26 |
| printf+string | 18.4 | 29 | 26 |
| IOStreams | 34.6 | 59 | 55 |
| fmt | 22.0 | 37 | 34 |
| tinyformat | 51.8 | 103 | 97 |
| Boost Format | 120.5 | 762 | 739 |
| Folly Format | 158.7 | 102 | 87 |
As you can see, fmt has 60% less overhead in terms of resulting binary code
size compared to IOStreams and comes pretty close to printf. Boost Format
has by far the largest overheads.
printf+string is the same as printf but with extra <string>
include to measure the overhead of the latter.
Non-optimized build
| Method | Compile Time, s | Executable size, KiB | Stripped size, KiB |
|---|---|---|---|
| printf | 2.4 | 33 | 30 |
| printf+string | 18.5 | 33 | 30 |
| IOStreams | 31.9 | 56 | 52 |
| fmt | 20.9 | 56 | 51 |
| tinyformat | 38.9 | 88 | 82 |
| Boost Format | 64.8 | 366 | 304 |
| Folly Format | 113.5 | 442 | 428 |
libc, lib(std)c++ and libfmt are all linked as shared
libraries to compare formatting function overhead only. Boost Format
and tinyformat are header-only libraries so they don't provide any
linkage options.
Please refer to Building the library for the instructions on how to build the library and run the unit tests.
Benchmarks reside in a separate repository, format-benchmarks, so to run the benchmarks you first need to clone this repository and generate Makefiles with CMake:
$ git clone --recursive https://github.com/fmtlib/format-benchmark.git $ cd format-benchmark $ cmake .
Then you can run the speed test:
$ make speed-test
or the bloat test:
$ make bloat-test
fmt is distributed under the BSD license.
The Format String Syntax section in the documentation is based on the one from Python string module documentation adapted for the current library. For this reason the documentation is distributed under the Python Software Foundation license available in doc/python-license.txt. It only applies if you distribute the documentation of fmt.
The fmt library is maintained by Victor Zverovich (vitaut) and Jonathan Müller (foonathan) with contributions from many other people. See Contributors and Releases for some of the names. Let us know if your contribution is not listed or mentioned incorrectly and we'll make it right.
The benchmark section of this readme file and the performance tests are taken from the excellent tinyformat library written by Chris Foster. Boost Format library is acknowledged transitively since it had some influence on tinyformat. Some ideas used in the implementation are borrowed from Loki SafeFormat and Diagnostic API in Clang. Format string syntax and the documentation are based on Python's str.format. Thanks Doug Turnbull for his valuable comments and contribution to the design of the type-safe API and Gregory Czajkowski for implementing binary formatting. Thanks Ruslan Baratov for comprehensive comparison of integer formatting algorithms and useful comments regarding performance, Boris Kaul for C++ counting digits benchmark. Thanks to CarterLi for contributing various improvements to the code.