Simple and minimalistic library that provides binary serialization in C
, C++
C#
and Java
.
All implementations support the same basic API described here, more information and documentation is provided for each implementation in its respective README file.
The library uses a Buffer
to serialize and deserialize data, the buffer holds the data written into it and can be set to write or read mode.
- once in read mode the buffer cannot be written into and only read operations may be performed.
- in write mode, only write operations will be possible. When in write mode the buffer may also be either dynamic or static.
- Dynamic Buffer - will adapt its size to the data being inserted into it.
- Static Buffer - has a set size that cannot change and trying to write into it over its size limit will result in an error.
- Creating a buffer:
- create a dynamic buffer in write mode.
- create a static buffer in write mode.
- create a buffer from byte array in write mode.
- Write into the buffer :
- Write primitive data types into the buffer.
- Write array types into the buffer.
- Write generic data types into the buffer, this means that any type can be serialized using a custom serializer, though some types are supported 'out of the box'.
- Read from buffer :
- Read primitive data types from a buffer.
- Read array types from a buffer, given array to read to and desired length.
- Read generic data types from a buffer. Note that usually a type being deserializable is a stronger constraint than being serializable and so a type might be serializable but not deserializable.
- Reading from a buffer will return the data that was read and will remove it from the buffer.
this is an example usage that uses pseudocode to demonstrate how the library is to be used:
Person {
id;
age;
// some more fiels..
}
person_serializer(person, buffer):
buffer.write(id);
buffer.write(age);
// write all the fields..
person_deserializer(person, buffer):
buffer.read(id);
buffer.read(age);
// read all the fields
...
// program that creates some data for a person object
Person person = aquire_person()
// now we want to save this person's data
Buffer buffer = make_buffer(type = dynamic)
buffer.write(person, person_serializer)
save_to_file(buffer.get_bytes())
...
// later we want to retrive the data and create a person
Buffer buffer = make_buffer(load_data())
buffer.set_read()
Person new_person() // empty person
buffer.read(new_person, person_deserializer)
//now new_person has the same state that person had before the serialization.
Information on how to build each implementation is provided in the respective README files, but for quick building the pack.py
script is provided to help build, test and consolidate all the binaries quickly and from one place, run ./pack.py -b
to rebuild all the implementations, this will create a release
folder containing folders with binaries for each implementation and a log folder with logs of the build and packaging.