[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Autoexpanding ZIP archives?

From: Michael Pilato <cmpilato_at_collab.net>
Date: 2005-12-08 18:41:57 CET

On Thu, 2005-12-08 at 11:19 -0600, Ben Collins-Sussman wrote:
> On 12/8/05, John Peacock <jpeacock@rowman.com> wrote:
>
> > You _can_ store your compressed source packages etc.,
> > but it is not any more efficient than storing it in a filesystem, and it
> > many ways it is less efficient...
>
> I don't know how we're less efficient. In the worst case, we take up
> just as much space as storing both versions of the file.
>
> IIRC, the binary diffs being stored in the repository are guaranteed
> never to be bigger than the the original file. So if you make a small
> change to a 10MB binary, and the resulting binary diff is >10MB, then
> I believe the repository just stores the fulltext instead of the diff.

Can't speak for the FSFS code, but the BDB code was certainly written in
this way.

But efficiency is measured in many ways. I suspect the "less efficient"
aspects John was talking about include the simple overhead of
interaction with a version control system at all, plus the cost of
binary decompression on access. Remember the motto, "Disk is cheap"?
Taken far enough, it somewhat devalues binary compression/deltication in
our back-end storage.

-- 
C. Michael Pilato <cmpilato@collab.net> 
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Received on Thu Dec 8 18:56:33 2005

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.