Excessive network traffic for pulling a small change

[Joke de Buhr]

On Donnerstag 11 März 2010 19:37:22 Colin D Bennett wrote:
> On Thu, 11 Mar 2010 19:27:58 +0100
> 
> Joke de Buhr <joke at seiken.de> wrote:
> > bzr+http:// and bzr+https:// are the smart server transports which
> > should be more efficient redarding the amount of transfered data.
> 
> Right.  But is the non-smart HTTP transport so inefficient that 2 MB+
> must be transferred to retrieve a revision containing only a few changed
> lines?

What exactly is transfered? One revision which causes only a few lines to be 
changed or a set of revisions which?

A set of revisions might actually result in a quiet large changeset.


rev: X
  bzr add file_10MB
  bzr commit -m "10 MB added"
rev: X+1
  bzr rm file_10MB
  bzr commit -m "file removed"
rev: X+2
  bzr commit -m "1 line added (less than 1kb)"


If the client is add revision X and pulls everything up to revision X+2 the 
10MB file needs to be transfered but since it's removed the within X+1 
"visible" change from X to X+1 would only be one line.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 835 bytes
Desc: This is a digitally signed message part.
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20100311/d4de7bbe/attachment.pgp