📝 overworked README

This commit is contained in:
Niels Lohmann 2017-01-24 14:37:25 +01:00
parent 250e5bf43d
commit daf8dcdb32

217
README.md
View file

@ -442,15 +442,20 @@ int vi = jn.get<int>();
// etc. // etc.
``` ```
### Arbitrary types conversions ### Arbitrary types conversions
Every type can be serialized in JSON, not just STL-containers and scalar types. Every type can be serialized in JSON, not just STL-containers and scalar types. Usually, you would do something along those lines:
Usually, you would do something along those lines:
```cpp ```cpp
namespace ns { namespace ns {
struct person { std::string name; std::string address; int age; }; struct person {
std::string name;
std::string address;
int age;
};
} }
// convert to JSON // convert to JSON
json j; json j;
ns::person p = createSomeone(); ns::person p = createSomeone();
@ -461,10 +466,14 @@ j["age"] = p.age;
// ... // ...
// convert from JSON // convert from JSON
ns::person p {j["name"].get<std::string>(), j["address"].get<std::string>(), j["age"].get<int>()}; ns::person p {
j["name"].get<std::string>(),
j["address"].get<std::string>(),
j["age"].get<int>()
};
``` ```
It works, but that's quite a lot of boilerplate.. Hopefully, there's a better way: It works, but that's quite a lot of boilerplate... Hopefully, there's a better way:
```cpp ```cpp
ns::person p = createPerson(); ns::person p = createPerson();
@ -482,27 +491,25 @@ To make this work with one of your types, you only need to provide two methods:
using nlohmann::json; using nlohmann::json;
namespace ns { namespace ns {
void to_json(json& j, person const& p) void to_json(json& j, person const& p) {
{ j = json{{"name", p.name}, {"address", p.address}, {"age", p.age}};
j = json{{"name", p.name}, {"address", p.address}, {"age", p.age}}; }
}
void from_json(json const& j, person& p) void from_json(json const& j, person& p) {
{ p.name = j["name"].get<std::string>();
p.name = j["name"].get<std::string>(); p.address = j["address"].get<std::string>();
p.address = j["address"].get<std::string>(); p.age = j["age"].get<int>();
p.age = j["age"].get<int>(); }
}
} // namespace ns } // namespace ns
``` ```
That's all. When calling the json constructor with your type, your custom `to_json` method will be automatically called. That's all! When calling the `json` constructor with your type, your custom `to_json` method will be automatically called.
Likewise, when calling `get<your_type>()`, the `from_json` method will be called. Likewise, when calling `get<your_type>()`, the `from_json` method will be called.
Some important things: Some important things:
* Those methods **MUST** be in your type's 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 in your type's namespace, or the library will not be able to locate them (in this example, they are in namespace `ns`, where `person` is defined).
* When using `get<your_type>()`, `your_type` **MUST** be DefaultConstructible and CopyConstructible (There is a way to bypass those requirements described later) * When using `get<your_type>()`, `your_type` **MUST** be [DefaultConstructible](http://en.cppreference.com/w/cpp/concept/DefaultConstructible) and [CopyConstructible](http://en.cppreference.com/w/cpp/concept/CopyConstructible). (There is a way to bypass those requirements described later.)
#### How do I convert third-party types? #### How do I convert third-party types?
@ -510,124 +517,117 @@ This requires a bit more advanced technique.
But first, let's see how this conversion mechanism works: But first, let's see how this conversion mechanism works:
The library uses **JSON Serializers** to convert types to json. 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](http://en.cppreference.com/w/cpp/language/adl)) The default serializer for `nlohmann::json` is `nlohmann::adl_serializer` (ADL means [Argument-Dependent Lookup](http://en.cppreference.com/w/cpp/language/adl)).
It is implemented like this (simplified): It is implemented like this (simplified):
```cpp ```cpp
template <typename T> template <typename T>
struct adl_serializer struct adl_serializer {
{ static void to_json(json& j, const T& value) {
static void to_json(json& j, const T& value) // calls the "to_json" method in T's namespace
{ }
// calls the "to_json" method in T's namespace
}
static void from_json(const json& j, T& value) static void from_json(const json& j, T& value) {
{ // same thing, but with the "from_json" method
// same thing, but with the "from_json" method }
}
}; };
``` ```
This serializer works fine when you have control over the type's namespace. 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`...
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: To solve this, you need to add a specialization of `adl_serializer` to the `nlohmann` namespace, here's an example:
```cpp ```cpp
// partial specialization (full specialization works too) // partial specialization (full specialization works too)
namespace nlohmann { namespace nlohmann {
template <typename T> template <typename T>
struct adl_serializer<boost::optional<T>> struct adl_serializer<boost::optional<T>> {
{ static void to_json(json& j, const boost::optional<T>& opt) {
static void to_json(json& j, const boost::optional<T>& opt) if (opt == boost::none) {
{ j = nullptr;
if (opt == boost::none) } else {
j = nullptr; j = *opt; // this will call adl_serializer<T>::to_json
else // which will find the free function to_json
j = *opt; // this will call adl_serializer<T>::to_json, which will find the free function to_json in T's namespace! // in T's namespace!
} }
}
static void from_json(const json& j, boost::optional<T>& opt) static void from_json(const json& j, boost::optional<T>& opt) {
{ if (!j.is_null()) {
if (!j.is_null()) opt = j.get<T>(); // same as above, but with
opt = j.get<T>(); // same as above, but with adl_serializer<T>::from_json // adl_serializer<T>::from_json
} }
}; }
};
} }
``` ```
#### How can I use `get()` for non-default constructible/non-copyable types? #### How can I use `get()` for non-default constructible/non-copyable types?
There is a way, if your type is **MoveConstructible**. There is a way, if your type is [MoveConstructible](http://en.cppreference.com/w/cpp/concept/MoveConstructible). You will need to specialize the `adl_serializer` as well, but with a special `from_json` overload:
You will need to specialize the `adl_serializer` as well, but with a special `from_json` overload:
```cpp ```cpp
struct move_only_type { struct move_only_type {
move_only_type() = delete; move_only_type() = delete;
move_only_type(int ii): i(ii) {} move_only_type(int ii): i(ii) {}
move_only_type(const move_only_type&) = delete; move_only_type(const move_only_type&) = delete;
move_only_type(move_only_type&&) = default; move_only_type(move_only_type&&) = default;
:
int i; private:
int i;
}; };
namespace nlohmann { namespace nlohmann {
template <> template <>
struct adl_serializer<move_only_type> struct adl_serializer<move_only_type> {
{ // note: the return type is no longer 'void',
// note: the return type is no longer 'void', and the method only takes one argument // and the method only takes one argument
static move_only_type from_json(const json& j) static move_only_type from_json(const json& j) {
{ return {j.get<int>()};
return {j.get<int>()}; }
}
// Here's the catch! You must provide a to_json method! // Here's the catch! You must provide a to_json method!
// Otherwise you will not be able to convert move_only_type to json, // Otherwise you will not be able to convert move_only_type to json,
// since you fully specialized adl_serializer on that type // since you fully specialized adl_serializer on that type
static void to_json(json& j, move_only_type t) static void to_json(json& j, move_only_type t) {
{ j = t.i;
j = t.i; }
} };
};
} }
``` ```
#### Can I write my own serializer? (Advanced use) #### 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. Yes. You might want to take a look at `[unit-udt.cpp](https://github.com/nlohmann/json/blob/develop/test/src/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: 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 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 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 * 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 <= 32, and uses ADL. Here is an example, without simplifications, that only accepts types with a size <= 32, and uses ADL.
```cpp ```cpp
// You should use void as a second template argument if you don't need compile-time checks on T // You should use void as a second template argument
template <typename T, typename SFINAE = typename std::enable_if<sizeof(T) <= 32>::type> // if you don't need compile-time checks on T
struct less_than_32_serializer // if someone tries to use a type bigger than 32, the compiler will complain template<typename T, typename SFINAE = typename std::enable_if<sizeof(T) <= 32>::type>
{ struct less_than_32_serializer {
template <typename Json> template <typename BasicJsonType>
static void to_json(Json& j, T value) static void to_json(BasicJsonType& j, T value) {
{ // we want to use ADL, and call the correct to_json overload
// we want to use ADL, and call the correct to_json overload using nlohmann::to_json; // this method is called by adl_serializer,
using nlohmann::to_json; // this method is called by adl_serializer, this is where the magic happens // this is where the magic happens
to_json(j, value); to_json(j, value);
} }
template <typename Json> template <typename BasicJsonType>
static void from_json(const Json& j, T& value) static void from_json(const BasicJsonType& j, T& value) {
{ // same thing here
// same thing here using nlohmann::from_json;
using nlohmann::from_json; from_json(j, value);
from_json(j, value); }
}
}; };
``` ```
@ -637,21 +637,19 @@ Be **very** careful when reimplementing your serializer, you can stack overflow
template <typename T, void> template <typename T, void>
struct bad_serializer struct bad_serializer
{ {
template <typename Json> template <typename BasicJsonType>
static void to_json(Json& j, const T& value) static void to_json(BasicJsonType& j, const T& value) {
{ // this calls BasicJsonType::json_serializer<T>::to_json(j, value);
// this calls Json::json_serializer<T>::to_json(j, value); // if BasicJsonType::json_serializer == bad_serializer ... oops!
// if Json::json_serializer == bad_serializer ... oops! j = value;
j = value; }
}
template <typename Json> template <typename BasicJsonType>
static void to_json(const Json& j, T& value) static void to_json(const BasicJsonType& j, T& value) {
{ // this calls BasicJsonType::json_serializer<T>::from_json(j, value);
// this calls Json::json_serializer<T>::from_json(j, value); // if BasicJsonType::json_serializer == bad_serializer ... oops!
// if Json::json_serializer == bad_serializer ... oops! value = j.template get<T>(); // oops!
value = j.template get<T>(); // oops! }
}
}; };
``` ```
@ -759,7 +757,7 @@ I deeply appreciate the help of the following people.
- [Eric Cornelius](https://github.com/EricMCornelius) pointed out a bug in the handling with NaN and infinity values. He also improved the performance of the string escaping. - [Eric Cornelius](https://github.com/EricMCornelius) pointed out a bug in the handling with NaN and infinity values. He also improved the performance of the string escaping.
- [易思龙](https://github.com/likebeta) implemented a conversion from anonymous enums. - [易思龙](https://github.com/likebeta) implemented a conversion from anonymous enums.
- [kepkin](https://github.com/kepkin) patiently pushed forward the support for Microsoft Visual studio. - [kepkin](https://github.com/kepkin) patiently pushed forward the support for Microsoft Visual studio.
- [gregmarr](https://github.com/gregmarr) simplified the implementation of reverse iterators and helped with numerous hints and improvements. - [gregmarr](https://github.com/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](https://github.com/caiovlp) fixed a bug in the Unicode handling. - [Caio Luppi](https://github.com/caiovlp) fixed a bug in the Unicode handling.
- [dariomt](https://github.com/dariomt) fixed some typos in the examples. - [dariomt](https://github.com/dariomt) fixed some typos in the examples.
- [Daniel Frey](https://github.com/d-frey) cleaned up some pointers and implemented exception-safe memory allocation. - [Daniel Frey](https://github.com/d-frey) cleaned up some pointers and implemented exception-safe memory allocation.
@ -787,7 +785,7 @@ I deeply appreciate the help of the following people.
- [duncanwerner](https://github.com/duncanwerner) found a really embarrassing performance regression in the 2.0.0 release. - [duncanwerner](https://github.com/duncanwerner) found a really embarrassing performance regression in the 2.0.0 release.
- [Damien](https://github.com/dtoma) fixed one of the last conversion warnings. - [Damien](https://github.com/dtoma) fixed one of the last conversion warnings.
- [Thomas Braun](https://github.com/t-b) fixed a warning in a test case. - [Thomas Braun](https://github.com/t-b) fixed a warning in a test case.
- [Théo DELRIEU](https://github.com/theodelrieu) patiently and constructively oversaw the long way toward [iterator-range parsing](https://github.com/nlohmann/json/issues/290). - [Théo DELRIEU](https://github.com/theodelrieu) patiently and constructively oversaw the long way toward [iterator-range parsing](https://github.com/nlohmann/json/issues/290). He also implemented the magic behind the serialization/deserialization of user-defined types.
- [Stefan](https://github.com/5tefan) fixed a minor issue in the documentation. - [Stefan](https://github.com/5tefan) fixed a minor issue in the documentation.
- [Vasil Dimov](https://github.com/vasild) fixed the documentation regarding conversions from `std::multiset`. - [Vasil Dimov](https://github.com/vasild) fixed the documentation regarding conversions from `std::multiset`.
- [ChristophJud](https://github.com/ChristophJud) overworked the CMake files to ease project inclusion. - [ChristophJud](https://github.com/ChristophJud) overworked the CMake files to ease project inclusion.
@ -801,6 +799,9 @@ I deeply appreciate the help of the following people.
- [Bosswestfalen](https://github.com/Bosswestfalen) merged two iterator classes into a smaller one. - [Bosswestfalen](https://github.com/Bosswestfalen) merged two iterator classes into a smaller one.
- [Daniel599](https://github.com/Daniel599) helped to get Travis execute the tests with Clang's sanitizers. - [Daniel599](https://github.com/Daniel599) helped to get Travis execute the tests with Clang's sanitizers.
- [Jonathan Lee](https://github.com/vjon) fixed an example in the README file. - [Jonathan Lee](https://github.com/vjon) fixed an example in the README file.
- [gnzlbg](https://github.com/gnzlbg) supported the implementation of user-defined types.
- [Alexej Harm](https://github.com/qis) helped to get the user-defined types working with Visual Studio.
- [Jared Grubb](https://github.com/jaredgrubb) supported the implementation of user-defined types.
Thanks a lot for helping out! Please [let me know](mailto:mail@nlohmann.me) if I forgot someone. Thanks a lot for helping out! Please [let me know](mailto:mail@nlohmann.me) if I forgot someone.