Author: Eugene Nalimov
Date: 15:10:39 02/11/02
Go up one level in this thread
Several years ago a friend of mine (Andrew Kadatch, authot of the compression/decompression code used fot the tablebases) reported a bug in the bzip2 to its author. Bzip2 can create the archieve it would not be able to extract. He never got the reply, and the bug is still there in the latest bzip2 sources. Eugene On February 11, 2002 at 17:47:52, Dann Corbit wrote: >On February 11, 2002 at 17:45:22, Ratko V Tomic wrote: > >>> Since bzip2 comes with full source code, it should be pretty >>> easy. Just include the decompressor and concatenate the >>> compressed data to the binary. >> >> >>The bzip2 offers about 30-50% better compression ratios (on color >>graphics) compared to gzip. But the public source implementation >>yields very uneven compression times, going from half the speed >>of gzip/zlib (at best) to 20 (and up) times slower for the >>"wrong" data. To make it work in real time one needs to >>pre-filter the data fed the bzip library so that its input >>has as few repeated character runs as possible. With such >>pre-filtering it works quite well. > >It might be nice to submit your improvements back to the open source project. I >am sure that they would be appreciated world-wide.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.