RFC: handlings large files via fragmenting

[Robert Collins]

On Mon, 2008-08-25 at 23:23 +0100, Adrian Wilkins wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> bazaar-request at lists.canonical.com wrote:
> > 
> > Message: 1
> > Date: Mon, 25 Aug 2008 13:02:21 -0400
> > From: Aaron Bentley <aaron at aaronbentley.com>
> 
> > Indeed, which is why I said "This requires a line-based delta approach".
> >  I'm well aware that groupcompress is not one.
> > 
> >
> 
> Am I being dense, or is a line-based delta less suitable for the kinds of files
> that are likely to be large? Bitmap graphics, databases, that sort of thing?
> About the only line-based format that I can imagine getting very large is
> something like the output of mysqldump.
> 
> Or do line-based deltas work equally well on binary files?

Many/most binary files have \n every 256 characters, on average. 

So line based deltas work quite well except on binary files that have a
very low \n occurrence rate.

However, a non-lined-based delta can do arbitrarily better at
compression.

-Rob
-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20080827/57fe68b4/attachment-0001.pgp