http://code.activestate.com/recipes/tags/rsync/new/2011-01-09T16:32:22-08:00ActiveState Code Recipes2011-01-09T16:32:22-08:00Eric Pruitthttp://code.activestate.com/recipes/users/4170757/http://code.activestate.com/recipes/577518-rsync-algorithm/

<p style="color: grey"> Python recipe 577518 by <a href="/recipes/users/4170757/">Eric Pruitt</a> (<a href="/recipes/tags/algorithm/">algorithm</a>, <a href="/recipes/tags/delta/">delta</a>, <a href="/recipes/tags/diff/">diff</a>, <a href="/recipes/tags/patch/">patch</a>, <a href="/recipes/tags/python3/">python3</a>, <a href="/recipes/tags/rsync/">rsync</a>). Revision 4. </p> <p>This is a pure Python implementation of the <a href="http://samba.anu.edu.au/rsync/">rsync algorithm</a>. On my desktop (3.0GHz dual core, 7200RPM), best case throughput for target file hash generation and delta generation is around 2.9MB/s. Absolute worst case scenario (no blocks in common) throughput for delta generation is 200KB/s to 300KB/s on the same system.</p> <p>Tested in Python 2.5, 2.6, and 3.1. In 2.7, io.BufferedReader should yield the best throughput. On all other versions use __builtin__.open.</p>