Version 1.1.0 (tag v1.1.0)

* Changed the code parameters according to the ones computed for tight
security bounds with the tools provided in the https://github.com/LEDAcrypt/LEDAtools
repository.

----------------------------
Version 1.0.2 (tag v1.0.2)

* Fixed decoding bug, unelicited by submission parameters.

* Fixed sparse multiplication bug, unelicited by submission parameters.

* Added forced zero filling for the memory area which subsequently holds 
  the data to be hashed whenever a decoding failure takes place.

----------------------------
Version 1.0.1 (tag v1.0.1)

* Added the secret key to the hash which generates the pseudorandom value
  for the encapsulated secret in case of decoding failure. This addresses 
  the official comment made by Keita Xagawa about the OW-CPA of the scheme.

* Added explicit count of the error vector weight. Decoding an error with
  a different weight from the expected one (t) always results in a decryption
  failure.
  Prevents trivial reaction attacks which forge messages with a number of errors
  (slightly) higher than the specification.

* Workaround for an apparent mistranslation of the code in the Clang/LLVM compilation
  toolchain of macOS Sierra 16.6.0, pointed out by Jacob Alperin-Sheriff.

----------------------------
Version 1.0 (tag v1.0)

* Original submission to NIST