2017-05-02:
There's yet another piece of faulty code in the key generation of (RE)BLISS.
Which caused signature generation failed to produce a valid signature.
2017-05-02:
As I said before, don't use the public-key part of my software just yet, they're experimental!
There's another possible piece of faulty code in the key generation of (RE)BLISS.
2016-09-20:
I generated the wrong Huffman coding table yet again!!
Don't use the public-key part of my software!!
2016-09-19:
!! Don't use the Public-Key part of my software !! Bugs are still being found !!
2016-08-09:
The 2016-08-08 release had a platform-specific Makefile in it, don't use it. And it's pkg-config version info is incorrect.
I made an publish script that'll handy these issues in the future.
2016-08-06:
!! Good News !! The AES block cipher and the GCM mode of operation has received optimization on x86 platforms, utilizing AES-NI and PCLMULQDQ instruction set extensions.
The latest version will be released with code has been tested.
2016-08-05:
After investigation, it was shown that the error was caused by allocating too short static buffers for the IV in testing code, not a bug in the implementation.
2016-08-05:
A bug has been noted that the implementation of GCM mode of operation fails for inivialization vectors longer than 12 bytes. Reason of which is under investigation.
2016-07-27:
!! URGENT !! UDATED !!
A critical bug has been noted while upgrading the code, that the infinity norm of z1 was not checked in BLISS.
The result of this bug is that all previous signatures are insecure and should be invalidated as soon as possible.
Since the suite generate Huffman coding table based on actual samples of signatures generated using DRBG instead of mathematical statistics, the bug has also invalidated the previous coding table.
2016-07-24:
The V.Singh RLWE-KEX will transition to NEWHOPE soon!
2016-07-09:
A bug has been noted in the mode-of-operation section of the suite.
Namely, AES counter cipher used for different key length were incorrectly using the AES-256 cipher function.
Right now, you can patch this error yourself, and I'm running every test I can find on the web for the correctness of the implementation. Once thoroughly tested, a new version will be released.
2017-09-13:
Fixed platform-independent endian conversion codes.
