Version 1.7.4, 2008-03-10
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

* Use unaligned memory read/writes on systems that allow it, for performance
* Assembly for x86-64 for accessing the bswap instruction
* Use larger buffers in ARC4 and WiderWAKE for significant throughput increase
* Unroll loops in SHA-160 for a few percent increase in performance
* Fix compilation with GCC 3.2 in es_ftw and es_unix
* Build fix for NetBSD systems
* Prevent es_dev from being built except on Unix systems