Re: WARNING! Object DB conversion (was Re: [PATCH] write-tree performance problems)

From: C. Scott Ananian <cscott@cscott.net>
Date: Apr 20, 2005, 14:35 UTC
Message-ID: <Pine.LNX.4.61.0504201031350.2630@cag.csail.mit.edu>
In-Reply-To: <Pine.LNX.4.58.0504200725110.6467@ppc970.osdl.org>

On Wed, 20 Apr 2005, Linus Torvalds wrote:

> - _keep_ the same compression format, but notice that we already have an
>   object by looking at the uncompressed one.

With a chunked file, you can also skip writing certain *subtrees* of the file as soon as you notice it's already present on disk. I can code this up if you are interested.

Of course, the paranoid folks will give up any performance benefit you 
obtain if they keep their "yes the SHA1 matches, but is the file *really* 
the same" code.  But maybe they're willing to be slow -- and they can do 
an uncompress rather than a compress in order to do the comparison, which 
will give *some* performance improvement.
  --scott

LCPANGS Serbian MKSEARCH security KUCLUB LCPANES Saddam Hussein Secretary 
Delta Force AMLASH ESMERALDITE TPAJAX plutonium ESGAIN Ft. Meade India
                          ( http://cscott.net/ )