Serializer

Library for binary packing and serialization. Might be useful when working with remote systems or embedded hardware.

Main classes

BinaryPacker and BinaryUnpacker

Use this class to pack and unpack data from byte vector. This class is not checking type safety and assumes you know what you're doing.

Simple usage:

std::vector<char> data;
gusc::Serializer::BinaryPacker packer(data);

packer << static_cast<uint8_t>(123) << true << 5;

gusc::Serializer::BinaryUnpacker unpacker(data);
uint8_t a;
unpacker >> a; // Should be 123
bool b;
unpacker >> b; // Should be true
int c;
unpacker >> c; // Should be 5

Extended usage with custom data types:

struct MyStruct {
    int a;
    bool b;
};

// This magic method allows to pack your custom data structure into byte stream
template<typename TChar>
void to_pack(gusc::Serializer::BinaryPacker<TChar>& packer, const MyStruct& inValue)
{
    packer << inValue.a << inValue.b;
}

// This magic method allows you to unpack yout custom data structure from byte stream
template<typename TChar>
void from_pack(gusc::Serializer::BinaryUnpacker<TChar>& unpacker, MyStruct& outValue)
{
    unpacker >> outValue.a;
    unpacker >> outValue.b;
}

// Simply continue using stream operators

MyStruct x = { 1, false };

std::vector<char> data;
gusc::Serializer::BinaryPacker packer(data);

packer << x; // Pack MyStruct into data stream

gusc::Serializer::BinaryUnpacker unpacker(data);

unpacker >> x; // Unpack MyStruct from data stream

BinarySerializer and BinaryDeserializer

Use this class to serialize and deserialize data from byte vector. This class is checking type safety and will throw a std::invalid_argument exception if requested type does not match.

Simple usage:

std::vector<char> data;
gusc::Serializer::BinarySerializer serializer(data);

serializer << static_cast<uint8_t>(123) << true << 5;

gusc::Serializer::BinaryDeserializer deserializer(data);
uint8_t a;
deserializer >> a; // Should be 123
bool b;
deserializer >> b; // Should be true
long c;
deserializer >> c; // will throw as we have written int instead of long

Extended usage with custom data types:

TODO