Performance improvements in Degu 0.7
Written for Degu by Jason Gerard DeRose on 2014-08-02
Degu 0.7 brings a C implementation of the `degu.base.
The new `read_preamble()` function is around 318% faster than the pure-Python equivalent in Degu 0.6.
As a result, benchmark.py is now around 20% faster for AF_INET6, and around 26% faster for AF_UNIX (on an Intel® Core™ i7-4900MQ when using the "performance" governor). Note that to reproduce this benchmark, you'll need to copy the benchmark.py script from the Degu 0.7 tree back into the Degu 0.6 tree.
Over time, all of `degu.base` will likely be replaced with a C extension, with a pure-Python reference implementation that can be used as a fallback when the C extension isn't available.
For more details, please see the 0.7 changelog:
http://
