Add NLOHMANN_DEFINE_DERIVED_TYPE_* macros

[rotolof]

I created the new macros:

• NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE
• NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE_WITH_DEFAULT
• NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE
• NLOHMANN_DEFINE_DERIVED_TYPE_NON_INTRUSIVE_WITH_DEFAULT

to recreate the proposed fix to #2199.
The macros accept as second parameter the base class.

Usage example:

class A {
  double Aa;
  double Ab;
  NLOHMANN_DEFINE_TYPE_INTRUSIVE(A, Aa, Ab)
};

class B : public A {
  int Ba;
  int Bb;
  NLOHMANN_DEFINE_DERIVED_TYPE_INTRUSIVE(B, A, Ba, Bb)
};

What they do is basically calling the to_json/from_json functions of the base class before serialising/deserialising the class members:

void to_json(nlohmann::json& j, const B& b) {
    nlohmann::to_json(j, static_cast<const A&>(b));
    // ...
}

void from_json(const nlohmann::json& j, B& b) {
    nlohmann::from_json(j, static_cast<A&>(b));
    // ...
}

---

Pull request checklist

Read the Contribution Guidelines for detailed information.

• Changes are described in the pull request, or an existing issue is referenced.
• The test suite compiles and runs without error.
• Code coverage is 100%. Test cases can be added by editing the test suite.
• The source code is amalgamated; that is, after making changes to the sources in the include/nlohmann directory, run make amalgamate to create the single-header files single_include/nlohmann/json.hpp and single_include/nlohmann/json_fwd.hpp. The whole process is described here.

[coveralls]

coverage: 100.0%. remained the same when pulling b6e2fd1 on rotolof:derived_macros into edffad0 on nlohmann:develop.

[hacker-cb]

Any updates about this PR?

[rotolof]

I just solved the conflict that had arisen in the meantime.

[nlohmann]

I am currently finishing #2998 to finally have a stable CI. Once this is merged, please update your branch.

[nlohmann]

Please add documentation. See https://github.com/nlohmann/json/blob/develop/docs/mkdocs/docs/api/macros/index.md for a starting point.

[nlohmann]

(Also, please update from the develop branch and fix the conflicts.)

[eyalk1]

since this code now actually modifies the json object from base - what do you guys think about using merge_patch or += instead of = for each member?
for example:

#define NLOHMANN_JSON_PAIR(member) {#member, nlohmann_json_t.member},

#define NLOHMANN_INIT_PAIRS(...)                                               \
  NLOHMANN_JSON_EXPAND(NLOHMANN_JSON_PASTE(NLOHMANN_JSON_PAIR, __VA_ARGS__))

#define NLOHMANN_DEFINE_TYPE_NON_INTRUSIVE_INHERITANCE(derived, base, ...)     \
  void to_json(nlohmann::json &nlohmann_json_j,                                \
               derived const &nlohmann_json_t) {                               \
    to_json(nlohmann_json_j, static_cast<base const &>(nlohmann_json_t));      \
    nlohmann_json_j.merge_patch(nlohmann::json{                                \
        NLOHMANN_JSON_EXPAND(NLOHMANN_INIT_PAIRS(__VA_ARGS__))});              \
  }                                                                            \
  void from_json(nlohmann::json const &nlohmann_json_j,                        \
                 derived &nlohmann_json_t) {                                   \
    from_json(nlohmann_json_j, static_cast<base &>(nlohmann_json_t));          \
    NLOHMANN_JSON_EXPAND(NLOHMANN_JSON_PASTE(NLOHMANN_JSON_FROM, __VA_ARGS__)) \
  }

[rotolof]

I updated to the latest develop and added documentation, let me know it is ok now

[github-actions]

🔴 Amalgamation check failed! 🔴

The source code has not been amalgamated. @rotolof
Please read and follow the Contribution Guidelines.

[github-actions]

🔴 Amalgamation check failed! 🔴

The source code has not been amalgamated. @rotolof
Please read and follow the Contribution Guidelines.

[gregmarr]

@rotolof Looks like you might be using a different version of astyle (perhaps even a newer one) which has different behavior than the current version being used. #4180 (comment)

[rotolof]

@gregmarr Thanks, I was using macOS brew's astyle that is 3.4.9. I tried with the suggested 3.1 but it does not support squeeze-lines, wtf?

[nlohmann]

Comment that line from the config and try again. Sorry for the inconvenience - I did not find the time to fix this yet.

[Aeshal]

Are there any chances to release that soon? Seems like the most of work has been done and it just hanging almost a year now.

[rotolof]

Are there any chances to release that soon? Seems like the most of work has been done and it just hanging almost a year now.

Sorry, I completely forgot about this until now! Let me check it again.

[Aeshal]

Is this change ok now? Can it possibly be integrated soon?

[rotolof]

Is this change ok now? Can it possibly be integrated soon?

I think that everything is fine, I had updated with the latest develop and all checks had passed, is anything missing @nlohmann?

[ncanderson]

Would love to see this released too.

[Aeshal]

@nlohmann can you please relate to this changes? Can we release that? It's ready from half a year now...

[nlohmann]

Looks good to me.

[nlohmann]

Thanks!