Serialize::load implementations now do sanity checks based on header information and return an error (ErrorKind::InvalidData) if the checks fail. This should catch invalid / corrupted data in many situations. Full validation would be too expensive for large data structures in high-performance applications.
Some redundant information has been removed from the header of SparseVector.
Serialize::loadimplementations now do sanity checks based on header information and return an error (ErrorKind::InvalidData) if the checks fail. This should catch invalid / corrupted data in many situations. Full validation would be too expensive for large data structures in high-performance applications.Some redundant information has been removed from the header of
SparseVector.