https://sourceforge.net/p/mhash/news/Recent posts to newsenFri, 14 Sep 2012 22:22:41 -0000https://sourceforge.net/p/mhash/news/2006/06/round-2-complete/Ok, mutils should now be more-or-less complete. Thread-unsafe functions are being fixed or eliminated. 64-bit calls will likely be added in future. Most of the redundant code has gone, but there is some still in there. Emphasis can now switch back to getting lots of new hash functions for people to play with.Jonathan DayFri, 14 Sep 2012 22:22:41 -0000https://sourceforge.netaa8792317d01deda9a4135d8fe29d9fb80a1b270https://sourceforge.net/p/mhash/news/2006/03/mhash---round-1-of-cleanup-complete/The first round of code cleanup and code speedup has been complete. Duplicate functions have been eliminated, type ranges are stricter, sanity checks have been added and a few (relatively minor) optimizations have been made. MHash now supports two numbering schemes for selecting hashes. The first is the standard scheme that has always been used. The second is a heirarchical scheme that - if it proves interesting - will be proposed to other crypto library writers as an interoperable convention. As always, code is welcome, patches are welcome, help is welcome, testing on strange and alien computer platforms from another galaxy is welcome.Jonathan DayFri, 14 Sep 2012 22:22:41 -0000https://sourceforge.net3befa831cd917f7c83326c0449184d3967cc5a95https://sourceforge.net/p/mhash/news/2004/04/mhash-update/New hashing functions - most notably Whirlpool, and the remainder of the SHA functions - have been added, along with RIPEMD. Documentation is lacking, but efforts to fix this are underway. There are only a few strong hashing functions not yet implemented in MHash. The addition of at least some of those, plus the docs and some code auditing will bring the version to 1.0.0. Examples include Cellhash, Parallel FFT-Hash and Subhash. The main efforts from 1.0.0 will be on speedups and security auditing. The idea here is that we want MHash to be not only the most comprehensive hashing library, but the most trustworthy as well. New algorithms'll still be added, if anything new or exciting comes along, provided there are grounds for saying it's useful, fast and secure. Faster implementations of existing algorithms are ALWAYS welcome.Jonathan DayFri, 14 Sep 2012 22:22:41 -0000https://sourceforge.net9befd9eab7fed5f07582c86705dfbbb80077f6ad