Re: [Boost-testing] Too many results?

From: Rene Rivera (grafikrobot_at_[hidden])
Date: 2007-09-27 15:02:03

Martin Wille wrote:
> Rene Rivera wrote:

>> So, any clever ideas out there?

> You could try (tar and) rzip instead of (tar and) bzip2. rzip uses the
> same compression algorithm as rsync. rzip is good in finding blocks of
> identical data even if these blocks are at locations far from each other
> in the to-be-compressed file. We've seen it do wonders on certain data.

What I'm afraid of is that the rsync algorithm is not good at 
compressing data with small widespread changes. The block matching 
algorithm isn't fine grained enough. Hence I'm skeptical that the idea 
of using rsync will help all that much. But I guess I need to do some 
experimentation to see how small of a diff set rsync can produce for the 
results.

Note, when I said ZIP I meant the real ZIP, not bzip, gzip, bzip2, etc. 
The ZIP archive is used directly by the web server. Something which is 
not possible with tar based archives.

-- 
-- Grafik - Don't Assume Anything
-- Redshift Software, Inc. - http://redshift-software.com
-- rrivera/acm.org - grafik/redshift-software.com
-- 102708583/icq - grafikrobot/aim - grafikrobot/yahoo