Differential D12544 Diff 41995 source/blender/blenlib/intern/serialize.cc

Changeset View

Standalone View

source/blender/blenlib/intern/serialize.cc

This file was added.

				#include "BLI_serialize.hh"

				#include "json.hpp"

				namespace blender::io::serialize {

				static void convert_to_json(nlohmann::json &j, const Value &value)
				{
				switch (value.type()) {
				case ValueType::String: {
				j = value.string_value();
				break;
				SeverinUnsubmitted Done Inline Actions Didn't the documentation for this function say that there will be an assert for the type-match as well? Severin: Didn't the documentation for this function say that there will be an assert for the type-match…
				jbakkerAuthorUnsubmitted Done Inline Actions Will adapt the documentation. seems like I had a different implementation in my head when writing the docs :-) jbakker: Will adapt the documentation. seems like I had a different implementation in my head when…
				}
				sybrenUnsubmitted Done Inline Actions This assumes that any `Value` with `type_ == eValueType::String` is a `StringValue`. This tight coupling is not documented at all! sybren: This assumes that any `Value` with `type_ == eValueType::String` is a `StringValue`. This tight…
				case ValueType::Int: {
				j = value.int_value();
				break;
				}
				}
				}

				void JsonFormatter::serialize(std::ostream &os, Value &value)
				{
				nlohmann::json j;
				convert_to_json(j, value);
				os << j.dump(2);
				}

				} // namespace blender::io::serialize
				SeverinUnsubmitted Done Inline Actions I would much prefer returning `std::unique_ptr`s here, rather than raw owning pointers. Especially since it's not documented that they are owning and need freeing :) Severin: I would much prefer returning `std::unique_ptr`s here, rather than raw owning pointers.
				sybrenUnsubmitted Done Inline Actions These should be extracted to their own functions. sybren: These should be extracted to their own functions.
				sybrenUnsubmitted Done Inline Actions Why is this necessary at all? sybren: Why is this necessary at all?
				sybrenUnsubmitted Done Inline Actions What is this notation? sybren: What is this notation?
				sybrenUnsubmitted Done Inline Actions Such blocks should be their own function sybren: Such blocks should be their own function
				sybrenUnsubmitted Done Inline Actions Then why are they (and this code) even here? sybren: Then why are they (and this code) even here?
				sybrenUnsubmitted Done Inline Actions will → would sybren: will → would
				sybrenUnsubmitted Done Inline Actions will → would sybren: will → would