C библиотека для JSON файлов в Windows

знаете ли вы библиотеки в C, которые управляют JSON файлами (в основном чтением и записью) в системе Windows?

Я не могу найти ни одной подходящей библиотеки, простой в использовании под windows, но я не эксперт!

Спасибо за любую поддержку, которую вы можете мне оказать.

2 ответа

Название довольно понятно: знает ли кто-нибудь (хорошую) библиотеку чтения файлов свойств для C или, если нет, C++? Edit: чтобы быть точным, мне нужна библиотека, которая обрабатывает формат файла .properties, используемый в Java: http://en.wikipedia.org/wiki/.свойства

Есть ли хорошая библиотека C/C++ для создания файлов PNG и которая поддерживает цветные полигоны и фигуры? Эта библиотека должна быть независимой от OS, так как она нужна мне как для Linux, так и для Windows.

JSON Домашняя страница содержит список всех библиотек для различных типов языков. Это то, что указано для C:

Ну, я могу найти хотя бы этот вариант .

Похожие вопросы:

Какие библиотеки, отличные от GPL, доступны для написания серверов и клиентов JSON-RPC на родном языке C++? Согласно http://json-rpc.org/wiki/implementations , по-видимому, существует только одна.

Я пытаюсь найти библиотеку для разбора JSON на C# на Windows Mobile (работаю с Visual Studio 2005). Библиотеки, которые я нашел, которые позволяют мне анализировать JSON в C# (litjson и Jayrock), не.

Возможный Дубликат : Разбор XML в чистом виде C Какая библиотека XML лучше всего подходит для C? Особенности, которые я ищу, таковы 1) Хорошая поддержка в виде учебных пособий и активного списка.

Название довольно понятно: знает ли кто-нибудь (хорошую) библиотеку чтения файлов свойств для C или, если нет, C++? Edit: чтобы быть точным, мне нужна библиотека, которая обрабатывает формат файла.

Есть ли хорошая библиотека C/C++ для создания файлов PNG и которая поддерживает цветные полигоны и фигуры? Эта библиотека должна быть независимой от OS, так как она нужна мне как для Linux, так и.

Есть ли разница между стандартной библиотекой и библиотекой времени выполнения? Является ли стандартная библиотека C в windows такой же, как и библиотека Windows C RT? Реализована ли она в.

В примечаниях к выпуску tensorflow 1.2 говорится, что Библиотека TensorFlow C теперь доступна для Windows. Ницца. Однако страница, посвященная C API на веб-сайте tensorflow, дает инструкции только.

Я ищу библиотеку C#, которая будет генерировать действительный объект JSON на основе заданной схемы JSON. Я хотел бы создать очень простой образец JSON точно так же, как это делает Swagger: Я видел.

какая библиотека лучше всего подходит для программирования windows сегодня в Visual C++, какая из них наиболее часто используется сегодня и т. д. ATL, MFC, WTL и другие.. а сколько их существует .

Я искал библиотеки JSON C++, и библиотека JSONcpp кажется наиболее используемой. Я хотел бы использовать для следующей задачи. В принципе, предположим, что у меня есть файл JSON, например:

Шпаргалка Java программиста 8. Библиотеки для работы с Json (Gson, Fastjson, LoganSquare, Jackson, JsonPath и другие)

В одной из моих прошлых статей я рассказывал о своем opensorce pet проекте useful-java-links, идея которого собрать как можно больше ссылок на полезные Java библиотеки и фреймворки. У него так же есть подпроект Hello World project идея которого для каждой библиотеки собрать несколько простых примеров её использования.

Проблема программистов в Java мире в том что кроме стандартной библиотеки JDK есть огромное других полезных библиотек, причем переход от одной библиотеки к другой может вызывать проблемы из-за неполной документации, отсутствия простых примеров или даже сложности понять какие зависимости нужно добавить в maven чтобы все запустилось. А на новой работе вполне могут использовать вместо твоей любимой библиотеки ту которую ты не знаешь. Идея моего проекта облегчить изучение и выбор разных библиотек.

Итак, давайте посмотрим какие известные библиотеки есть для работы с JSON в Java…

8. Работа с Json

JSON парсеры

Аналог XPath для JSON

Генерация Java классов из JSON или JSON схемы и JSON валидация

Итак, у нас восемь библиотек для сериализации и десериализации в json, две библиотеки для генерации Java классов по схеме или json файлу, одна библиотека для валидации схемы и два аналога XPath, но для json. Давайте рассмотрим каждую из них.

1. JSON парсеры

Существует три основных способа сериализации и десериализации среди указанных библиотек (от самого простого к самому сложному) и один дополнительный:

1. Data bind,
2. Tree Model,
3. Streaming API,
4. (И дополнительный способ) Аналоги XPath,

Давайте рассмотрим с чем их едят:

Data bind самый популярный и простой способ, вы просто указываете класс, который нужно преобразовать в json, может быть часть полей отмечаете аннотациями (а зачастую даже это необязательно), а библиотека сама превращает этот класс и всю его иерархию классов в json. Аналогом при работе с xml будет JAXB (Java Architecture for XML Binding)
Плюсы: наиболее простой из всех, по сути главное реализовать только Java классы, более того можно просто сгенерировать Java классы из json’a или json схемы.
Минусы: скорость и память. Большинство библиотек использует рефлексию и т.п. методы работы с Java классами (хотя не все), что очевидно не очень быстро. К тому же, весь json файл сразу превращается в Java объекты, что может просто исчерпать всю доступную память, если вы попытаетесь обработать очень большой json.
Вывод: если нет проблем с производительностью, памятью и вы не собираетесь обрабатывать многогигабайтные json’ы скорее всего самый лучший способ.

Tree Model — данный парсер представляет json в виде Java классов таких как Node или JsonElement c иерархической структурой, а уже сам программист их обходит и получает из них информацию. Данный способ похож на DOM парсеры в xml.
Плюсы: обычно быстрее первого способа и проще третьего,
Минусы: уступает Data bind по простоте, плюс ряд библиотек способен генерить классы при Data bind, а не использовать рефлексию, в этом случае то что Tree Model будет быстрее не очевидно, к тому же не решается проблема огромных файлов и ограничения памяти.

Streaming API — самый низкоуровневый способ, по сути программист сам вручную разбирает токены json’a. Зато никаких ограничений по памяти и в теории максимальная производительность.
Плюсы: производительность и минимальное потребление памяти,
Минусы: сложность использования,

Аналоги XPath — дополнительный способ, не очень подходит, если нужно получит всю информацию из json’a, зато позволяет написав выражение $.store.book[*].author и получить список всех авторов всех книг из json’a магазина. То есть легко получать часть информации из json’а.
Плюсы: позволяет быстро получить информацию из json’а по сложным критериям,
Минусы: не очень подходит, когда нужна все информация из json’а, не работает в обратную сторону на формирования json’ов,

1.1 Обзор библиотек

Способ	Fastjson	Gson	LoganSquare	JSON java	Moshi	Ig json parser	Jackson	Genson	JsonPath
1. Data bind	Да	Да	Да	—	Да	Да	Да	Да	—
2. Tree Model	—	Да	—	Да	—	—	Да	—	—
3. Streaming API	—	Да	—	—	—	—	Да	—	—
4. Аналоги XPath	Да	—	—	—	—	—	—	—	Да
5. Генерация классов для Data bind*	—	—	Да	—	—	Да	—	—	—
6. Github’s star	4851	4120	2188	1937	1732	921	881	108	849
7. Работает со static inner class**	Да	Да	Нет	—	Да	Нет	Да	Да	—
8. Обязательность аннотаций***	Нет	Нет	Да	—	Нет	Да	Нет	Нет	—

По ссылкам на Да можно найти примеры использования.
* — Генерация классов для Data bind позволяет сгенерировать классы на стадии компиляции, что в теории должно давать значительный прирост производительности библиотеки,
** — Работает со static inner class имеет смысл только для случая Data bind, возможно ли сериализация и десериализация для случая статических внутренних классов (не статические внутренние классы сериализовать не рекомендуется),
*** — тоже только для случая Data bind можно ли не использовать аннотации или их использование крайне рекомендуется,

1.2 Простейшие примеры использование Data bind

Для демонстрации работы библиотек будем использовать следующий json:

И следующие Java классы (в разных примерах могут слегка отличаться наличием аннотаций, если они обязательны):

Как можно увидеть, Java классы всего лишь состоять из двух классов Human и Place, в которых храниться сообщение Hi World. Json тоже содержит эти два вложенных объекта.

Примеры использования (Data bind): Способ	Fastjson	Gson	LoganSquare	Moshi	Ig json parser	Jackson	Genson
Инициализация	—	Gson gson = new Gson()	—	Moshi moshi = new Moshi. Builder().build(); JsonAdapter jsonAdapter = moshi.adapter(Human.class)	—	ObjectMapper mapper = new ObjectMapper()	Genson genson = new Genson()
Из Java в json	JSON.toJSONString(human)	gson.toJson(human)	LoganSquare.serialize(human)	jsonAdapter.toJson(human)	Human__JsonHelper.serializeToJson(human)	mapper.writeValueAsString(human)	genson.serialize(human)
Из json в Java	JSON.parseObject(jsonString, Human.class)	gson.fromJson(jsonString, Human.class)	LoganSquare.parse(jsonString, Human.class)	jsonAdapter.fromJson(jsonString)	Human__JsonHelper.parseFromJson(jsonString)	mapper.readValue(jsonString, Human.class)	genson.deserialize(jsonString, Human.class)

Human__JsonHelper — это класс который Ig json parser сгенерировал на этапе компиляции, у LoganSquare так же есть генерации на этапе компиляции, но там классы подключаются «под капотом» внутри LoganSquare.

libjson

Brought to you by: ninja9578

Downloads: 200 This Week

Last Update: 2016-11-09

A JSON reader and writer which is super-effiecient and usually runs circles around other JSON libraries. It’s highly customizable to optimize for your particular project, and very lightweight. For Windows, OSX, or Linux. Works in any language.

Features

• Lazy or aggressive JSON Parsing
• 100% JSON compliant
• Language independent C interface
• C++ interface
• Test Suite and Example Projects
• C and Bash style comment support
• Automated install via make
• cplusplus.com — style documentation
• Streaming ability
• Security to prevent various forms of Denial of Service

Project Activity

Follow libjson

Other Useful Business Software

Software that helps organizations gain actionable insights from customers, employees, citizens, users, and attendees.

Rate This Project

User Ratings

User Reviews

MacOS required some corrections and changes. Line 2 of GNU_C.h fixed misspelling of _GUN_ to _GNU_. From within _internal/Sources directory: ln -s ../Dependencies. make; sudo make install Expect warnings: clang: warning: optimization flag ‘-fexpensive-optimizations’ is not supported clang: warning: argument unused during compilation: ‘-fast’ clang: warning: argument unused during compilation: ‘-fexpensive-optimizations’ g++ -o $@ $

Works as advertised, easy to use.

Be careful when you are up to choose this libjson. 1. If you need to handle Unicode ( why not. You have to for handling Twitter & Facebook response. ), make sure that your host app uses GNU C++ standard not strict ANSI C++ standard. if using strict ANSI C++ standard, Unicode support in this library should be turned off. Then it can’t handle Unicode escape sequence like \uXXXX. At least it should return \uXXXX if it can’t handle the Unicode. As some people pointed out, wouldn’t it be better to conver it to UTF-8? Unicode escape sequence is not UTF-8. If Unicode option is off, it will return chopped string at \u. (returns only up to \. eg. «This is \uAC00» yields «This \». 2. Documentation is not clear if it’s SOX style or DOM parser style. It turned out that it handles the both, which is good. More accurately speaking, you can achieve SOX-like on by using its iterator. 3. Its build option is also expected to be specified in a header file, JSONOptions.h, which can introduce some odd situation when using this library in your host project.

I had to make a change to the make file because of a «No such file or directory». After the change, it worked like a charm. The change is in the install_headers: banner section. cp -rv ./$(srcdir)/Dependencies/ $(include_path)/$(libname_hdr)/$(srcdir) was changed to cp -rv ./$(srcdir)/../Dependencies/ $(include_path)/$(libname_hdr)/$(srcdir)

Json библиотека для windows

There are myriads of JSON libraries out there, and each may even have its reason to exist. Our class had these design goals:

Intuitive syntax. In languages such as Python, JSON feels like a first class data type. We used all the operator magic of modern C++ to achieve the same feeling in your code. Check out the examples below and you’ll know what I mean.

Trivial integration. Our whole code consists of a single header file json.hpp . That’s it. No library, no subproject, no dependencies, no complex build system. The class is written in vanilla C++11. All in all, everything should require no adjustment of your compiler flags or project settings.

Serious testing. Our class is heavily unit-tested and covers 100% of the code, including all exceptional behavior. Furthermore, we checked with Valgrind and the Clang Sanitizers that there are no memory leaks. Google OSS-Fuzz additionally runs fuzz tests against all parsers 24/7, effectively executing billions of tests so far. To maintain high quality, the project is following the Core Infrastructure Initiative (CII) best practices.

Other aspects were not so important to us:

Memory efficiency. Each JSON object has an overhead of one pointer (the maximal size of a union) and one enumeration element (1 byte). The default generalization uses the following C++ data types: std::string for strings, int64_t , uint64_t or double for numbers, std::map for objects, std::vector for arrays, and bool for Booleans. However, you can template the generalized class basic_json to your needs.

Speed. There are certainly faster JSON libraries out there. However, if your goal is to speed up your development by adding JSON support with a single header, then this library is the way to go. If you know how to use a std::vector or std::map , you are already set.

See the contribution guidelines for more information.

You can sponsor this library at GitHub Sponsors.

🏷️ Named Sponsors

json.hpp is the single required file in single_include/nlohmann or released here. You need to add

to the files you want to process JSON and set the necessary switches to enable C++11 (e.g., -std=c++11 for GCC and Clang).

You can further use file include/nlohmann/json_fwd.hpp for forward-declarations. The installation of json_fwd.hpp (as part of cmake’s install step), can be achieved by setting -DJSON_MultipleHeaders=ON .

You can also use the nlohmann_json::nlohmann_json interface target in CMake. This target populates the appropriate usage requirements for INTERFACE_INCLUDE_DIRECTORIES to point to the appropriate include directories and INTERFACE_COMPILE_FEATURES for the necessary C++11 flags.

To use this library from a CMake project, you can locate it directly with find_package() and use the namespaced imported target from the generated package configuration:

The package configuration file, nlohmann_jsonConfig.cmake , can be used either from an install tree or directly out of the build tree.

To embed the library directly into an existing CMake project, place the entire source tree in a subdirectory and call add_subdirectory() in your CMakeLists.txt file:

Since CMake v3.11, FetchContent can be used to automatically download the repository as a dependency at configure type.

Note: The repository https://github.com/nlohmann/json download size is huge. It contains all the dataset used for the benchmarks. You might want to depend on a smaller repository. For instance, you might want to replace the URL above by https://github.com/ArthurSonzogni/nlohmann_json_cmake_fetchcontent

To allow your project to support either an externally supplied or an embedded JSON library, you can use a pattern akin to the following:

thirdparty/nlohmann_json is then a complete copy of this source tree.

🍺 If you are using OS X and Homebrew, just type brew tap nlohmann/json and brew install nlohmann-json and you’re set. If you want the bleeding edge rather than the latest release, use brew install nlohmann-json —HEAD .

If you are using the Meson Build System, add this source tree as a meson subproject. You may also use the include.zip published in this project’s Releases to reduce the size of the vendored source tree. Alternatively, you can get a wrap file by downloading it from Meson WrapDB, or simply use meson wrap install nlohmann_json . Please see the meson project for any issues regarding the packaging.

The provided meson.build can also be used as an alternative to cmake for installing nlohmann_json system-wide in which case a pkg-config file is installed. To use it, simply have your build system require the nlohmann_json pkg-config dependency. In Meson, it is preferred to use the dependency() object with a subproject fallback, rather than using the subproject directly.

If you are using Conan to manage your dependencies, merely add nlohmann_json/x.y.z to your conanfile ‘s requires, where x.y.z is the release version you want to use. Please file issues here if you experience problems with the packages.

If you are using Spack to manage your dependencies, you can use the nlohmann-json package. Please see the spack project for any issues regarding the packaging.

If you are using hunter on your project for external dependencies, then you can use the nlohmann_json package. Please see the hunter project for any issues regarding the packaging.

If you are using Buckaroo, you can install this library’s module with buckaroo add github.com/buckaroo-pm/nlohmann-json . Please file issues here. There is a demo repo here.

If you are using vcpkg on your project for external dependencies, then you can use the nlohmann-json package. Please see the vcpkg project for any issues regarding the packaging.

If you are using cget, you can install the latest development version with cget install nlohmann/json . A specific version can be installed with cget install nlohmann/json@v3.1.0 . Also, the multiple header version can be installed by adding the -DJSON_MultipleHeaders=ON flag (i.e., cget install nlohmann/json -DJSON_MultipleHeaders=ON ).

If you are using CocoaPods, you can use the library by adding pod «nlohmann_json», ‘

>3.1.2′ to your podfile (see an example). Please file issues here.

If you are using NuGet, you can use the package nlohmann.json. Please check this extensive description on how to use the package. Please files issues here.

If you are using conda, you can use the package nlohmann_json from conda-forge executing conda install -c conda-forge nlohmann_json . Please file issues here.

If you are using MSYS2, your can use the mingw-w64-nlohmann-json package, just type pacman -S mingw-w64-i686-nlohmann-json or pacman -S mingw-w64-x86_64-nlohmann-json for installation. Please file issues here if you experience problems with the packages.

If you are using build2 , you can use the nlohmann-json package from the public repository https://cppget.org or directly from the package’s sources repository. In your project’s manifest file, just add depends: nlohmann-json (probably with some version constraints). If you are not familiar with using dependencies in build2 , please read this introduction. Please file issues here if you experience problems with the packages.

If you are using wsjcpp , you can use the command wsjcpp install «https://github.com/nlohmann/json:develop» to get the latest version. Note you can change the branch «:develop» to an existing tag or another branch.

If you are using CPM.cmake , you can check this example . After adding CPM script to your project, implement the following snippet to your CMake:

If you are using bare Makefiles, you can use pkg-config to generate the include flags that point to where the library is installed:

Users of the Meson build system will also be able to use a system wide library, which will be found by pkg-config :

Beside the examples below, you may want to check the documentation where each function contains a separate code example (e.g., check out emplace() ). All example files can be compiled and executed on their own (e.g., file emplace.cpp).

JSON as first-class data type

Here are some examples to give you an idea how to use the class.

Assume you want to create the JSON object

With this library, you could write:

Note that in all these cases, you never need to «tell» the compiler which JSON value type you want to use. If you want to be explicit or express some edge cases, the functions json::array() and json::object() will help:

You can create a JSON value (deserialization) by appending _json to a string literal:

Note that without appending the _json suffix, the passed string literal is not parsed, but just used as JSON string value. That is, json j = «< \"happy\": true, \"pi\": 3.141 >» would just store the string «< "happy": true, "pi": 3.141 >» rather than parsing the actual object.

The above example can also be expressed explicitly using json::parse() :

You can also get a string representation of a JSON value (serialize):

Note the difference between serialization and assignment:

.dump() returns the originally stored string value.

Note the library only supports UTF-8. When you store strings with different encodings in the library, calling dump() may throw an exception unless json::error_handler_t::replace or json::error_handler_t::ignore are used as error handlers.

To/from streams (e.g. files, string streams)

You can also use streams to serialize and deserialize:

These operators work for any subclasses of std::istream or std::ostream . Here is the same example with files:

Please note that setting the exception bit for failbit is inappropriate for this use case. It will result in program termination due to the noexcept specifier in use.

Read from iterator range

You can also parse JSON from an iterator range; that is, from any container accessible by iterators whose value_type is an integral type of 1, 2 or 4 bytes, which will be interpreted as UTF-8, UTF-16 and UTF-32 respectively. For instance, a std::vector , or a std::list :

You may leave the iterators for the range [begin, end):

Custom data source

Since the parse function accepts arbitrary iterator ranges, you can provide your own data sources by implementing the LegacyInputIterator concept.

The library uses a SAX-like interface with the following functions:

The return value of each function determines whether parsing should proceed.

To implement your own SAX handler, proceed as follows:

1. Implement the SAX interface in a class. You can use class nlohmann::json_sax as base class, but you can also use any class where the functions described above are implemented and public.
2. Create an object of your SAX interface class, e.g. my_sax .
3. Call bool json::sax_parse(input, &my_sax) ; where the first parameter can be any input like a string or an input stream and the second parameter is a pointer to your SAX interface.

Note the sax_parse function only returns a bool indicating the result of the last executed SAX event. It does not return a json value — it is up to you to decide what to do with the SAX events. Furthermore, no exceptions are thrown in case of a parse error — it is up to you what to do with the exception object passed to your parse_error implementation. Internally, the SAX interface is used for the DOM parser (class json_sax_dom_parser ) as well as the acceptor ( json_sax_acceptor ), see file json_sax.hpp .

We designed the JSON class to behave just like an STL container. In fact, it satisfies the ReversibleContainer requirement.

Conversion from STL containers

Any sequence container ( std::array , std::vector , std::deque , std::forward_list , std::list ) whose values can be used to construct JSON values (e.g., integers, floating point numbers, Booleans, string types, or again STL containers described in this section) can be used to create a JSON array. The same holds for similar associative containers ( std::set , std::multiset , std::unordered_set , std::unordered_multiset ), but in these cases the order of the elements of the array depends on how the elements are ordered in the respective STL container.

Likewise, any associative key-value containers ( std::map , std::multimap , std::unordered_map , std::unordered_multimap ) whose keys can construct an std::string and whose values can be used to construct JSON values (see examples above) can be used to create a JSON object. Note that in case of multimaps only one key is used in the JSON object and the value depends on the internal order of the STL container.

JSON Pointer and JSON Patch

The library supports JSON Pointer (RFC 6901) as alternative means to address structured values. On top of this, JSON Patch (RFC 6902) allows to describe differences between two JSON values — effectively allowing patch and diff operations known from Unix.

JSON Merge Patch

The library supports JSON Merge Patch (RFC 7386) as a patch format. Instead of using JSON Pointer (see above) to specify values to be manipulated, it describes the changes using a syntax that closely mimics the document being modified.

Supported types can be implicitly converted to JSON values.

It is recommended to NOT USE implicit conversions FROM a JSON value. You can find more details about this recommendation here. You can switch off implicit conversions by defining JSON_USE_IMPLICIT_CONVERSIONS to 0 before including the json.hpp header. When using CMake, you can also achieve this by setting the option JSON_ImplicitConversions to OFF .

Note that char types are not automatically converted to JSON strings, but to integer numbers. A conversion to a string must be specified explicitly:

Arbitrary types conversions

Every type can be serialized in JSON, not just STL containers and scalar types. Usually, you would do something along those lines:

It works, but that’s quite a lot of boilerplate. Fortunately, there’s a better way:

To make this work with one of your types, you only need to provide two functions:

That’s all! When calling the json constructor with your type, your custom to_json method will be automatically called. Likewise, when calling get () or get_to(your_type&) , the from_json method will be called.

Some important things:

• Those methods MUST be in your type’s namespace (which can be the global namespace), or the library will not be able to locate them (in this example, they are in namespace ns , where person is defined).
• Those methods MUST be available (e.g., proper headers must be included) everywhere you use these conversions. Look at issue 1108 for errors that may occur otherwise.
• When using get () , your_type MUST be DefaultConstructible. (There is a way to bypass this requirement described later.)
• In function from_json , use function at() to access the object values rather than operator[] . In case a key does not exist, at throws an exception that you can handle, whereas operator[] exhibits undefined behavior.
• You do not need to add serializers or deserializers for STL types like std::vector : the library already implements these.

Simplify your life with macros

If you just want to serialize/deserialize some structs, the to_json / from_json functions can be a lot of boilerplate.

There are two macros to make your life easier as long as you (1) want to use a JSON object as serialization and (2) want to use the member variable names as object keys in that object:

• NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE(name, member1, member2, . ) is to be defined inside of the namespace of the class/struct to create code for.
• NLOHMANN_DEFINE_TYPE_INTRUSIVE(name, member1, member2, . ) is to be defined inside of the class/struct to create code for. This macro can also access private members.

In both macros, the first parameter is the name of the class/struct, and all remaining parameters name the members.

The to_json / from_json functions for the person struct above can be created with:

Here is an example with private members, where NLOHMANN_DEFINE_TYPE_INTRUSIVE is needed:

How do I convert third-party types?

This requires a bit more advanced technique. But first, let’s see how this conversion mechanism works:

The library uses JSON Serializers to convert types to json. The default serializer for nlohmann::json is nlohmann::adl_serializer (ADL means Argument-Dependent Lookup).

It is implemented like this (simplified):

This serializer works fine when you have control over the type’s namespace. However, what about boost::optional or std::filesystem::path (C++17)? Hijacking the boost namespace is pretty bad, and it’s illegal to add something other than template specializations to std .

To solve this, you need to add a specialization of adl_serializer to the nlohmann namespace, here’s an example:

How can I use get() for non-default constructible/non-copyable types?

There is a way, if your type is MoveConstructible. You will need to specialize the adl_serializer as well, but with a special from_json overload:

Can I write my own serializer? (Advanced use)

Yes. You might want to take a look at unit-udt.cpp in the test suite, to see a few examples.

If you write your own serializer, you’ll need to do a few things:

• use a different basic_json alias than nlohmann::json (the last template parameter of basic_json is the JSONSerializer )
• use your basic_json alias (or a template parameter) in all your to_json / from_json methods
• use nlohmann::to_json and nlohmann::from_json when you need ADL

Here is an example, without simplifications, that only accepts types with a size

Specializing enum conversion

By default, enum values are serialized to JSON as integers. In some cases this could result in undesired behavior. If an enum is modified or re-ordered after data has been serialized to JSON, the later de-serialized JSON data may be undefined or a different enum value than was originally intended.

It is possible to more precisely specify how a given enum is mapped to and from JSON as shown below:

The NLOHMANN_JSON_SERIALIZE_ENUM() macro declares a set of to_json() / from_json() functions for type TaskState while avoiding repetition and boilerplate serialization code.

Usage:

• NLOHMANN_JSON_SERIALIZE_ENUM() MUST be declared in your enum type’s namespace (which can be the global namespace), or the library will not be able to locate it and it will default to integer serialization.
• It MUST be available (e.g., proper headers must be included) everywhere you use the conversions.

Other Important points:

• When using get () , undefined JSON values will default to the first pair specified in your map. Select this default pair carefully.
• If an enum or JSON value is specified more than once in your map, the first matching occurrence from the top of the map will be returned when converting to or from JSON.

Binary formats (BSON, CBOR, MessagePack, and UBJSON)

Though JSON is a ubiquitous data format, it is not a very compact format suitable for data exchange, for instance over a network. Hence, the library supports BSON (Binary JSON), CBOR (Concise Binary Object Representation), MessagePack, and UBJSON (Universal Binary JSON Specification) to efficiently encode JSON values to byte vectors and to decode such vectors.

The library also supports binary types from BSON, CBOR (byte strings), and MessagePack (bin, ext, fixext). They are stored by default as std::vector to be processed outside of the library.

Though it’s 2021 already, the support for C++11 is still a bit sparse. Currently, the following compilers are known to work:

• GCC 4.8 — 11.0 (and possibly later)
• Clang 3.4 — 11.0 (and possibly later)
• Apple Clang 9.1 — 12.3 (and possibly later)
• Intel C++ Compiler 17.0.2 (and possibly later)
• Microsoft Visual C++ 2015 / Build Tools 14.0.25123.0 (and possibly later)
• Microsoft Visual C++ 2017 / Build Tools 15.5.180.51428 (and possibly later)
• Microsoft Visual C++ 2019 / Build Tools 16.3.1+1def00d3d (and possibly later)

I would be happy to learn about other compilers/versions.

GCC 4.8 has a bug 57824): multiline raw strings cannot be the arguments to macros. Don’t use multiline raw strings directly in macros with this compiler.

Android defaults to using very old compilers and C++ libraries. To fix this, add the following to your Application.mk . This will switch to the LLVM C++ library, the Clang compiler, and enable C++11 and other features disabled by default.

The code compiles successfully with Android NDK, Revision 9 — 11 (and possibly later) and CrystaX’s Android NDK version 10.

For GCC running on MinGW or Android SDK, the error ‘to_string’ is not a member of ‘std’ (or similarly, for strtod or strtof ) may occur. Note this is not an issue with the code, but rather with the compiler itself. On Android, see above to build with a newer environment. For MinGW, please refer to this site and this discussion for information on how to fix this bug. For Android NDK using APP_STL := gnustl_static , please refer to this discussion.

Unsupported versions of GCC and Clang are rejected by #error directives. This can be switched off by defining JSON_SKIP_UNSUPPORTED_COMPILER_CHECK . Note that you can expect no support in this case.

The following compilers are currently used in continuous integration at Travis, AppVeyor, and GitHub Actions:

Compiler	Operating System	CI Provider
Apple Clang 10.0.1 (clang-1001.0.46.4); Xcode 10.2.1	macOS 10.14.4	Travis
Apple Clang 10.0.1 (clang-1001.0.46.4); Xcode 10.3	macOS 10.15.7	GitHub Actions
Apple Clang 11.0.0 (clang-1100.0.33.12); Xcode 11.2.1	macOS 10.15.7	GitHub Actions
Apple Clang 11.0.0 (clang-1100.0.33.17); Xcode 11.3.1	macOS 10.15.7	GitHub Actions
Apple Clang 11.0.3 (clang-1103.0.32.59); Xcode 11.4.1	macOS 10.15.7	GitHub Actions
Apple Clang 11.0.3 (clang-1103.0.32.62); Xcode 11.5	macOS 10.15.7	GitHub Actions
Apple Clang 11.0.3 (clang-1103.0.32.62); Xcode 11.6	macOS 10.15.7	GitHub Actions
Apple Clang 11.0.3 (clang-1103.0.32.62); Xcode 11.7	macOS 10.15.7	GitHub Actions
Apple Clang 12.0.0 (clang-1200.0.32.2); Xcode 12	macOS 10.15.7	GitHub Actions
Apple Clang 12.0.0 (clang-1200.0.32.21); Xcode 12.1	macOS 10.15.7	GitHub Actions
Apple Clang 12.0.0 (clang-1200.0.32.21); Xcode 12.1.1	macOS 10.15.7	GitHub Actions
Apple Clang 12.0.0 (clang-1200.0.32.27); Xcode 12.2	macOS 10.15.7	GitHub Actions
Apple Clang 12.0.0 (clang-1200.0.32.28); Xcode 12.3	macOS 10.15.7	GitHub Actions
Apple Clang 12.0.0 (clang-1200.0.32.29); Xcode 12.4	macOS 10.15.7	GitHub Actions
GCC 4.8.5 (Ubuntu 4.8.5-4ubuntu2)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 4.9.3 (Ubuntu 4.9.3-13ubuntu2)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 5.4.0 (Ubuntu 5.4.0-6ubuntu1 16.04.12)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 6.5.0 (Ubuntu 6.5.0-2ubuntu1 14.04.1)	Ubuntu 14.04.5 LTS	Travis
GCC 7.5.0 (Ubuntu 7.5.0-6ubuntu2)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 8.1.0 (x86_64-posix-seh-rev0, Built by MinGW-W64 project)	Windows-10.0.17763	GitHub Actions
GCC 8.1.0 (i686-posix-dwarf-rev0, Built by MinGW-W64 project)	Windows-10.0.17763	GitHub Actions
GCC 8.4.0 (Ubuntu 8.4.0-3ubuntu2)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 9.3.0 (Ubuntu 9.3.0-17ubuntu1 20.04)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 10.2.0 (Ubuntu 10.2.0-5ubuntu1 20.04)	Ubuntu 20.04.2 LTS	GitHub Actions
GCC 11.0.1 20210321 (experimental)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 3.5.2 (3.5.2-3ubuntu1)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 3.6.2 (3.6.2-3ubuntu2)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 3.7.1 (3.7.1-2ubuntu2)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 3.8.0 (3.8.0-2ubuntu4)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 3.9.1 (3.9.1-4ubuntu3 16.04.2)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 4.0.0 (4.0.0-1ubuntu1 16.04.2)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 5.0.0 (5.0.0-3 16.04.1)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 6.0.1 (6.0.1-14)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 7.0.1 (7.0.1-12)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 8.0.1 (8.0.1-9)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 9.0.1 (9.0.1-12)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 10.0.0 (10.0.0-4ubuntu1)	Ubuntu 20.04.2 LTS	GitHub Actions
Clang 10.0.0 with GNU-like command-line	Windows-10.0.17763	GitHub Actions
Clang 11.0.0 with GNU-like command-line	Windows-10.0.17763	GitHub Actions
Clang 11.0.0 with MSVC-like command-line	Windows-10.0.17763	GitHub Actions
Clang 11.1.0 (11.1.0-++20210204121720+1fdec59bffc1-1 exp1 20210203232336.162	Ubuntu 20.04.2 LTS	GitHub Actions
Visual Studio 14 2015 MSVC 19.0.24241.7 (Build Engine version 14.0.25420.1)	Windows-6.3.9600	AppVeyor
Visual Studio 15 2017 MSVC 19.16.27035.0 (Build Engine version 15.9.21+g9802d43bc3 for .NET Framework)	Windows-10.0.14393	AppVeyor
Visual Studio 15 2017 MSVC 19.16.27045.0 (Build Engine version 15.9.21+g9802d43bc3 for .NET Framework)	Windows-10.0.14393	GitHub Actions
Visual Studio 16 2019 MSVC 19.28.29912.0 (Build Engine version 16.9.0+57a23d249 for .NET Framework)	Windows-10.0.17763	GitHub Actions
Visual Studio 16 2019 MSVC 19.28.29912.0 (Build Engine version 16.9.0+57a23d249 for .NET Framework)	Windows-10.0.17763	AppVeyor

The class is licensed under the MIT License:

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

The class contains the UTF-8 Decoder from Bjoern Hoehrmann which is licensed under the MIT License (see above). Copyright © 2008-2009 Björn Hoehrmann bjoern@hoehrmann.de

The class contains a slightly modified version of the Grisu2 algorithm from Florian Loitsch which is licensed under the MIT License (see above). Copyright © 2009 Florian Loitsch

The class contains a copy of Hedley from Evan Nemerson which is licensed as CC0-1.0.

The class contains parts of Google Abseil which is licensed under the Apache 2.0 License.

If you have questions regarding the library, I would like to invite you to open an issue at GitHub. Please describe your request, problem, or question as detailed as possible, and also mention the version of the library you are using as well as the version of your compiler and operating system. Opening an issue at GitHub allows other users and contributors to this library to collaborate. For instance, I have little experience with MSVC, and most issues in this regard have been solved by a growing community. If you have a look at the closed issues, you will see that we react quite timely in most cases.

Only if your request would contain confidential information, please send me an email. For encrypted messages, please use this key.

I deeply appreciate the help of the following people.

• Teemperor implemented CMake support and lcov integration, realized escape and Unicode handling in the string parser, and fixed the JSON serialization.
• elliotgoodrich fixed an issue with double deletion in the iterator classes.
• kirkshoop made the iterators of the class composable to other libraries.
• wancw fixed a bug that hindered the class to compile with Clang.
• Tomas Åblad found a bug in the iterator implementation.
• Joshua C. Randall fixed a bug in the floating-point serialization.
• Aaron Burghardt implemented code to parse streams incrementally. Furthermore, he greatly improved the parser class by allowing the definition of a filter function to discard undesired elements while parsing.
• Daniel Kopeček fixed a bug in the compilation with GCC 5.0.
• Florian Weber fixed a bug in and improved the performance of the comparison operators.
• Eric Cornelius pointed out a bug in the handling with NaN and infinity values. He also improved the performance of the string escaping.
• 易思龙 implemented a conversion from anonymous enums.
• kepkin patiently pushed forward the support for Microsoft Visual studio.
• gregmarr simplified the implementation of reverse iterators and helped with numerous hints and improvements. In particular, he pushed forward the implementation of user-defined types.
• Caio Luppi fixed a bug in the Unicode handling.
• dariomt fixed some typos in the examples.
• Daniel Frey cleaned up some pointers and implemented exception-safe memory allocation.
• Colin Hirsch took care of a small namespace issue.
• Huu Nguyen correct a variable name in the documentation.
• Silverweed overloaded parse() to accept an rvalue reference.
• dariomt fixed a subtlety in MSVC type support and implemented the get_ref() function to get a reference to stored values.
• ZahlGraf added a workaround that allows compilation using Android NDK.
• whackashoe replaced a function that was marked as unsafe by Visual Studio.
• 406345 fixed two small warnings.
• Glen Fernandes noted a potential portability problem in the has_mapped_type function.
• Corbin Hughes fixed some typos in the contribution guidelines.
• twelsby fixed the array subscript operator, an issue that failed the MSVC build, and floating-point parsing/dumping. He further added support for unsigned integer numbers and implemented better roundtrip support for parsed numbers.
• Volker Diels-Grabsch fixed a link in the README file.
• msm- added support for American Fuzzy Lop.
• Annihil fixed an example in the README file.
• Themercee noted a wrong URL in the README file.
• Lv Zheng fixed a namespace issue with int64_t and uint64_t .
• abc100m analyzed the issues with GCC 4.8 and proposed a partial solution.
• zewt added useful notes to the README file about Android.
• Róbert Márki added a fix to use move iterators and improved the integration via CMake.
• Chris Kitching cleaned up the CMake files.
• Tom Needham fixed a subtle bug with MSVC 2015 which was also proposed by Michael K..
• Mário Feroldi fixed a small typo.
• duncanwerner found a really embarrassing performance regression in the 2.0.0 release.
• Damien fixed one of the last conversion warnings.
• Thomas Braun fixed a warning in a test case and adjusted MSVC calls in the CI.
• Théo DELRIEU patiently and constructively oversaw the long way toward iterator-range parsing. He also implemented the magic behind the serialization/deserialization of user-defined types and split the single header file into smaller chunks.
• Stefan fixed a minor issue in the documentation.
• Vasil Dimov fixed the documentation regarding conversions from std::multiset .
• ChristophJud overworked the CMake files to ease project inclusion.
• Vladimir Petrigo made a SFINAE hack more readable and added Visual Studio 17 to the build matrix.
• Denis Andrejew fixed a grammar issue in the README file.
• Pierre-Antoine Lacaze found a subtle bug in the dump() function.
• TurpentineDistillery pointed to std::locale::classic() to avoid too much locale joggling, found some nice performance improvements in the parser, improved the benchmarking code, and realized locale-independent number parsing and printing.
• cgzones had an idea how to fix the Coverity scan.
• Jared Grubb silenced a nasty documentation warning.
• Yixin Zhang fixed an integer overflow check.
• Bosswestfalen merged two iterator classes into a smaller one.
• Daniel599 helped to get Travis execute the tests with Clang’s sanitizers.
• Jonathan Lee fixed an example in the README file.
• gnzlbg supported the implementation of user-defined types.
• Alexej Harm helped to get the user-defined types working with Visual Studio.
• Jared Grubb supported the implementation of user-defined types.
• EnricoBilla noted a typo in an example.
• Martin Hořeňovský found a way for a 2x speedup for the compilation time of the test suite.
• ukhegg found proposed an improvement for the examples section.
• rswanson-ihi noted a typo in the README.
• Mihai Stan fixed a bug in the comparison with nullptr s.
• Tushar Maheshwari added cotire support to speed up the compilation.
• TedLyngmo noted a typo in the README, removed unnecessary bit arithmetic, and fixed some -Weffc++ warnings.
• Krzysztof Woś made exceptions more visible.
• ftillier fixed a compiler warning.
• tinloaf made sure all pushed warnings are properly popped.
• Fytch found a bug in the documentation.
• Jay Sistar implemented a Meson build description.
• Henry Lee fixed a warning in ICC and improved the iterator implementation.
• Vincent Thiery maintains a package for the Conan package manager.
• Steffen fixed a potential issue with MSVC and std::min .
• Mike Tzou fixed some typos.
• amrcode noted a misleading documentation about comparison of floats.
• Oleg Endo reduced the memory consumption by replacing with .
• dan-42 cleaned up the CMake files to simplify including/reusing of the library.
• Nikita Ofitserov allowed for moving values from initializer lists.
• Greg Hurrell fixed a typo.
• Dmitry Kukovinets fixed a typo.
• kbthomp1 fixed an issue related to the Intel OSX compiler.
• Markus Werle fixed a typo.
• WebProdPP fixed a subtle error in a precondition check.
• Alex noted an error in a code sample.
• Tom de Geus reported some warnings with ICC and helped fixing them.
• Perry Kundert simplified reading from input streams.
• Sonu Lohani fixed a small compilation error.
• Jamie Seward fixed all MSVC warnings.
• Nate Vargas added a Doxygen tag file.
• pvleuven helped fixing a warning in ICC.
• Pavel helped fixing some warnings in MSVC.
• Jamie Seward avoided unnecessary string copies in find() and count() .
• Mitja fixed some typos.
• Jorrit Wronski updated the Hunter package links.
• Matthias Möller added a .natvis for the MSVC debug view.
• bogemic fixed some C++17 deprecation warnings.
• Eren Okka fixed some MSVC warnings.
• abolz integrated the Grisu2 algorithm for proper floating-point formatting, allowing more roundtrip checks to succeed.
• Vadim Evard fixed a Markdown issue in the README.
• zerodefect fixed a compiler warning.
• Kert allowed to template the string type in the serialization and added the possibility to override the exceptional behavior.
• mark-99 helped fixing an ICC error.
• Patrik Huber fixed links in the README file.
• johnfb found a bug in the implementation of CBOR’s indefinite length strings.
• Paul Fultz II added a note on the cget package manager.
• Wilson Lin made the integration section of the README more concise.
• RalfBielig detected and fixed a memory leak in the parser callback.
• agrianius allowed to dump JSON to an alternative string type.
• Kevin Tonon overworked the C++11 compiler checks in CMake.
• Axel Huebl simplified a CMake check and added support for the Spack package manager.
• Carlos O’Ryan fixed a typo.
• James Upjohn fixed a version number in the compilers section.
• Chuck Atkins adjusted the CMake files to the CMake packaging guidelines and provided documentation for the CMake integration.
• Jan Schöppach fixed a typo.
• martin-mfg fixed a typo.
• Matthias Möller removed the dependency from std::stringstream .
• agrianius added code to use alternative string implementations.
• Daniel599 allowed to use more algorithms with the items() function.
• Julius Rakow fixed the Meson include directory and fixed the links to cppreference.com.
• Sonu Lohani fixed the compilation with MSVC 2015 in debug mode.
• grembo fixed the test suite and re-enabled several test cases.
• Hyeon Kim introduced the macro JSON_INTERNAL_CATCH to control the exception handling inside the library.
• thyu fixed a compiler warning.
• David Guthrie fixed a subtle compilation error with Clang 3.4.2.
• Dennis Fischer allowed to call find_package without installing the library.
• Hyeon Kim fixed an issue with a double macro definition.
• Ben Berman made some error messages more understandable.
• zakalibit fixed a compilation problem with the Intel C++ compiler.
• mandreyel fixed a compilation problem.
• Kostiantyn Ponomarenko added version and license information to the Meson build file.
• Henry Schreiner added support for GCC 4.8.
• knilch made sure the test suite does not stall when run in the wrong directory.
• Antonio Borondo fixed an MSVC 2017 warning.
• Dan Gendreau implemented the NLOHMANN_JSON_SERIALIZE_ENUM macro to quickly define a enum/JSON mapping.
• efp added line and column information to parse errors.
• julian-becker added BSON support.
• Pratik Chowdhury added support for structured bindings.
• David Avedissian added support for Clang 5.0.1 (PS4 version).
• Jonathan Dumaresq implemented an input adapter to read from FILE* .
• kjpus fixed a link in the documentation.
• Manvendra Singh fixed a typo in the documentation.
• ziggurat29 fixed an MSVC warning.
• Sylvain Corlay added code to avoid an issue with MSVC.
• mefyl fixed a bug when JSON was parsed from an input stream.
• Millian Poquet allowed to install the library via Meson.
• Michael Behrns-Miller found an issue with a missing namespace.
• Nasztanovics Ferenc fixed a compilation issue with libc 2.12.
• Andreas Schwab fixed the endian conversion.
• Mark-Dunning fixed a warning in MSVC.
• Gareth Sylvester-Bradley added operator/ for JSON Pointers.
• John-Mark noted a missing header.
• Vitaly Zaitsev fixed compilation with GCC 9.0.
• Laurent Stacul fixed compilation with GCC 9.0.
• Ivor Wanders helped reducing the CMake requirement to version 3.1.
• njlr updated the Buckaroo instructions.
• Lion fixed a compilation issue with GCC 7 on CentOS.
• Isaac Nickaein improved the integer serialization performance and implemented the contains() function.
• past-due suppressed an unfixable warning.
• Elvis Oric improved Meson support.
• Matěj Plch fixed an example in the README.
• Mark Beckwith fixed a typo.
• scinart fixed bug in the serializer.
• Patrick Boettcher implemented push_back() and pop_back() for JSON Pointers.
• Bruno Oliveira added support for Conda.
• Michele Caini fixed links in the README.
• Hani documented how to install the library with NuGet.
• Mark Beckwith fixed a typo.
• yann-morin-1998 helped reducing the CMake requirement to version 3.1.
• Konstantin Podsvirov maintains a package for the MSYS2 software distro.
• remyabel added GNUInstallDirs to the CMake files.
• Taylor Howard fixed a unit test.
• Gabe Ron implemented the to_string method.
• Watal M. Iwasaki fixed a Clang warning.
• Viktor Kirilov switched the unit tests from Catch to doctest
• Juncheng E fixed a typo.
• tete17 fixed a bug in the contains function.
• Xav83 fixed some cppcheck warnings.
• 0xflotus fixed some typos.
• Christian Deneke added a const version of json_pointer::back .
• Julien Hamaide made the items() function work with custom string types.
• Evan Nemerson updated fixed a bug in Hedley and updated this library accordingly.
• Florian Pigorsch fixed a lot of typos.
• Camille Bégué fixed an issue in the conversion from std::pair and std::tuple to json .
• Anthony VH fixed a compile error in an enum deserialization.
• Yuriy Vountesmery noted a subtle bug in a preprocessor check.
• Chen fixed numerous issues in the library.
• Antony Kellermann added a CI step for GCC 10.1.
• Alex fixed an MSVC warning.
• Rainer proposed an improvement in the floating-point serialization in CBOR.
• Francois Chabot made performance improvements in the input adapters.
• Arthur Sonzogni documented how the library can be included via FetchContent .
• Rimas Misevičius fixed an error message.
• Alexander Myasnikov fixed some examples and a link in the README.
• Hubert Chathi made CMake’s version config file architecture-independent.
• OmnipotentEntity implemented the binary values for CBOR, MessagePack, BSON, and UBJSON.
• ArtemSarmini fixed a compilation issue with GCC 10 and fixed a leak.
• Evgenii Sopov integrated the library to the wsjcpp package manager.
• Sergey Linev fixed a compiler warning.
• Miguel Magalhães fixed the year in the copyright.
• Gareth Sylvester-Bradley fixed a compilation issue with MSVC.
• Alexander “weej” Jones fixed an example in the README.
• Antoine Cœur fixed some typos in the documentation.
• jothepro updated links to the Hunter package.
• Dave Lee fixed link in the README.
• Joël Lamotte added instruction for using Build2’s package manager.
• Paul Jurczak fixed an example in the README.
• Sonu Lohani fixed a warning.
• Carlos Gomes Martinho updated the Conan package source.
• Konstantin Podsvirov fixed the MSYS2 package documentation.
• Tridacnid improved the CMake tests.
• Michael fixed MSVC warnings.
• Quentin Barbarat fixed an example in the documentation.
• XyFreak fixed a compiler warning.
• TotalCaesar659 fixed links in the README.
• Tanuj Garg improved the fuzzer coverage for UBSAN input.
• AODQ fixed a compiler warning.
• jwittbrodt made NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE inline.
• pfeatherstone improved the upper bound of arguments of the NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE / NLOHMANN_DEFINE_TYPE_INTRUSIVE macros.
• Jan Procházka fixed a bug in the CBOR parser for binary and string values.
• T0b1-iOS fixed a bug in the new hash implementation.
• Matthew Bauer adjusted the CBOR writer to create tags for binary subtypes.
• gatopeich implemented an ordered map container for nlohmann::ordered_json .
• Érico Nogueira Rolim added support for pkg-config.
• KonanM proposed an implementation for the NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE / NLOHMANN_DEFINE_TYPE_INTRUSIVE macros.
• Guillaume Racicot implemented string_view support and allowed C++20 support.
• Alex Reinking improved CMake support for FetchContent .
• Hannes Domani provided a GDB pretty printer.

Thanks a lot for helping out! Please let me know if I forgot someone.

Used third-party tools

The library itself consists of a single header file licensed under the MIT license. However, it is built, tested, documented, and whatnot using a lot of third-party tools and services. Thanks a lot!

Projects using JSON for Modern C++

The library is currently used in Apple macOS Sierra and iOS 10. I am not sure what they are using the library for, but I am happy that it runs on so many devices.

The library supports Unicode input as follows:

• Only UTF-8 encoded input is supported which is the default encoding for JSON according to RFC 8259.
• std::u16string and std::u32string can be parsed, assuming UTF-16 and UTF-32 encoding, respectively. These encodings are not supported when reading from files or other input containers.
• Other encodings such as Latin-1 or ISO 8859-1 are not supported and will yield parse or serialization errors.
• Unicode noncharacters will not be replaced by the library.
• Invalid surrogates (e.g., incomplete pairs such as \uDEAD ) will yield parse errors.
• The strings stored in the library are UTF-8 encoded. When using the default string type ( std::string ), note that its length/size functions return the number of stored bytes rather than the number of characters or glyphs.
• When you store strings with different encodings in the library, calling dump() may throw an exception unless json::error_handler_t::replace or json::error_handler_t::ignore are used as error handlers.

Comments in JSON

This library does not support comments by default. It does so for three reasons:

Comments are not part of the JSON specification. You may argue that // or /* */ are allowed in JavaScript, but JSON is not JavaScript.

This was not an oversight: Douglas Crockford wrote on this in May 2012:

I removed comments from JSON because I saw people were using them to hold parsing directives, a practice which would have destroyed interoperability. I know that the lack of comments makes some people sad, but it shouldn’t.

Suppose you are using JSON to keep configuration files, which you would like to annotate. Go ahead and insert all the comments you like. Then pipe it through JSMin before handing it to your JSON parser.

It is dangerous for interoperability if some libraries would add comment support while others don’t. Please check The Harmful Consequences of the Robustness Principle on this.

However, you can pass set parameter ignore_comments to true in the parse function to ignore // or /* */ comments. Comments will then be treated as whitespace.

Order of object keys

By default, the library does not preserve the insertion order of object elements. This is standards-compliant, as the JSON standard defines objects as «an unordered collection of zero or more name/value pairs».

If you do want to preserve the insertion order, you can try the type nlohmann::ordered_json . Alternatively, you can use a more sophisticated ordered map like tsl::ordered_map (integration) or nlohmann::fifo_map (integration).

We checked with Valgrind and the Address Sanitizer (ASAN) that there are no memory leaks.

If you find that a parsing program with this library does not release memory, please consider the following case and it maybe unrelated to this library.

Your program is compiled with glibc. There is a tunable threshold that glibc uses to decide whether to actually return memory to the system or whether to cache it for later reuse. If in your program you make lots of small allocations and those small allocations are not a contiguous block and are presumably below the threshold, then they will not get returned to the OS. Here is a related issue #1924.

• The code contains numerous debug assertions which can be switched off by defining the preprocessor macro NDEBUG , see the documentation of assert . In particular, note operator[] implements unchecked access for const objects: If the given key is not present, the behavior is undefined (think of a dereferenced null pointer) and yields an assertion failure if assertions are switched on. If you are not sure whether an element in an object exists, use checked access with the at() function. Furthermore, you can define JSON_ASSERT(x) to replace calls to assert(x) .
• As the exact type of a number is not defined in the JSON specification, this library tries to choose the best fitting C++ number type automatically. As a result, the type double may be used to store numbers which may yield floating-point exceptions in certain rare situations if floating-point exceptions have been unmasked in the calling code. These exceptions are not caused by the library and need to be fixed in the calling code, such as by re-masking the exceptions prior to calling library functions.
• The code can be compiled without C++ runtime type identification features; that is, you can use the -fno-rtti compiler flag.
• Exceptions are used widely within the library. They can, however, be switched off with either using the compiler flag -fno-exceptions or by defining the symbol JSON_NOEXCEPTION . In this case, exceptions are replaced by abort() calls. You can further control this behavior by defining JSON_THROW_USER (overriding throw ), JSON_TRY_USER (overriding try ), and JSON_CATCH_USER (overriding catch ). Note that JSON_THROW_USER should leave the current scope (e.g., by throwing or aborting), as continuing after it may yield undefined behavior.

Execute unit tests

To compile and run the tests, you need to execute

Note that during the ctest stage, several JSON test files are downloaded from an external repository. If policies forbid downloading artifacts during testing, you can download the files yourself and pass the directory with the test files via -DJSON_TestDataDirectory=path to CMake. Then, no Internet connectivity is required. See issue #2189 for more information.

In case you have downloaded the library rather than checked out the code via Git, test cmake_fetch_content_configure . Please execute ctest -LE git_required to skip these tests. See issue #2189 for more information.

Some tests change the installed files and hence make the whole process not reproducible. Please execute ctest -LE not_reproducible to skip these tests. See issue #2324 for more information.

Note you need to call cmake -LE «not_reproducible|git_required» to exclude both labels. See issue #2596 for more information.

As Intel compilers use unsafe floating point optimization by default, the unit tests may fail. Use flag /fp:precise then.