Ser20 is a C++20 fork of cereal, a C++11 serialization library. It is completely compatible except for a namespace change. Thanks to modern meta-programming, Ser20 is 15% smaller and can compile up to 25% faster. Ser20 also has some other runtime optimizations, and is optimized for debug experience and debug symbols size.
Ser20 requires compiler support for C++20 concepts. Clang 13 and GCC10 should be modern enough, but CI testing is not set up very hermetically yet.
The best way to use Ser20 is with CMake. Ser20 is not header-only, which means that a compilation step is necessary to use it. The simplest way to do that is to use Ser20 as a subproject in your CMake project.
Using Ser20 is pretty simple. Except for a new namespace, it is completely compatible with Ser20 and therefor pretty compatible with boost::serialize as well.
This is a simple usage example:
#include <ser20/types/unordered_map.hpp>
#include <ser20/types/memory.hpp>
#include <ser20/archives/binary.hpp>
#include <fstream>
struct MyRecord {
uint8_t x, y;
float z;
template <class Archive> void serialize(Archive& ar) { ar(x, y, z); }
};
struct SomeData {
int32_t id;
std::shared_ptr<std::unordered_map<uint32_t, MyRecord>> data;
template <class Archive> void save(Archive& ar) const { ar(data); }
template <class Archive> void load(Archive& ar) {
static int32_t idGen = 0;
id = idGen++;
ar(data);
}
};
int main() {
std::ofstream os("out.bin", std::ios::binary);
Ser20::BinaryOutputArchive archive(os);
SomeData myData;
archive(myData);
return 0;
}Although it's by far not the fastest serialization library out there, Ser20 is sometimes two times as fast as cereal and boost::serialization in binary serialization:
This difference is mostly due to buffering of small reads and writes to ostreams.
Ser20 is licensed under the BSD license.
