Cista Versions Save

New features:

• Serializable high-performance hash map and hash set implementation based on Google's Swiss Table technique. By default, hash and equality are generated automatically.
• Optional feature: Automatic data structure versioning based on generated type-hashes
• Optional feature: checksum to prevent deserialization of corrupt data (data integrity check)
• Memory mapped file target for serialization (~6x speed-up for practical serialization test)
• Big endian support
• More ways to index values (indexed_vector<T>, indexed<T>) to be able to point to them
• Updated documentation

Development: fuzzing (also integrated into CI)

cista::offset and cista::raw contain functions and data structures for offset based and raw serialization

offset_ptr: stores the offset as difference between this and the object it points to. Serializing this allows to skip the deserialization step and just use a reinterpret_cast<T>(buf.begin()). This is useful for example in situations with shared memory.

Initial basic serialization functionality. For more details look into README.md.