Some unscientific timing results (on the Python source tree)

[Robert Collins]

On Mon, 2008-03-24 at 12:16 +0000, Paul Moore wrote:

> (Just a clarification - I've been toying with both Mercurial and
> Bazaar for my personal work for a number of months now. Mercurial's
> speed and flexibility are strong plus points for me, but overall I
> prefer Bazaar's UI, feature set, and development pace. I've watched
> Bazaar drive for more speed with great interest, as it would
> immediately swing the decision for me if there were no speed
> difference, but I'm no longer sure, now that packs have landed, if
> there's the same drive for speed as there was in the run-up to 1.0).

Well, perhaps I can answer that :).

packs have delivered a storage layer we believe we can scale with no
fundamental changes for a long time. *some* code paths have been
optimised to work with the storage layer, and those ones are the ones
that are very snappy. There is a lot of fat left to remove from bzr
(without changing the exterior of the system), to fully take advantage
of packs.

Right now we're working on:
 - log
 - shrinking the repository size (by up to 1/2)
 - reducing the time to extract texts from the repository
 - increasing streaming performance
 - allowing stacked branches, which like cvs/svn will store most or all
   of their history remotely, allowing casual contributors to download
   much less data - and importantly upload much less data when they 
   publish a branch

-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/20080325/d839bae0/attachment.pgp