[MERGE][RFC] further add performance improvements

[Robert Collins]

On Fri, 2006-05-19 at 09:22 -0500, John Arbash Meinel wrote:
> Robert Collins wrote:
> > here is a draft patch for another 100% improvement over what Martin and
> > I did this afternoon.
> > 
> > This patch refactors smart_add_tree to use the parent inventory entry
> > rather than looking up the path at every step along the tree. This gives
> > me a profiled performance of 5s for me on a 10K file tree, down from 10s
> > with the work we did earlier today, or 25s before tuning at all.
> > 
> > Its a little ugly because it has to allow the dry run mode to carry on
> > working, which is actually broken I think anyhow. Perhaps a 'dry run'
> > flag is actually better here? (That could trigger a
> > tree._read_inventory() rather than a write_inventory, for instance).
> > This would mean that the callbacks to add could always add, which means
> > we'd be able to get at the tree information more simply because the
> > parent would be guaranteed once we moved out of the user supplied files.
> > 
> > Rob
> > 
> 
> I think we could have 'add' always add the files to the inventory. We
> just don't have to write the inventory to the disk when we are done.

That was my example above - not calling write_inventory. Sounds like you
agreed.

Do I remember you doing some profiling on inventory writing performance
at some point ?

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: 191 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060520/8194d17f/attachment.pgp