post 0.8 development

[Robert Collins]

On Wed, 2006-05-10 at 12:38 +1000, Martin Pool wrote:
> On  8 May 2006, John A Meinel <john at arbash-meinel.com> wrote:

>  
> > The problem is that the round-trip overhead is probably murderous. You 
> > have a 'list_dir()' call for each directory, and then a 'stat' call for 
> > every entry (to determine if it is a file or a directory).
> 
> Perhaps a reasonable compromise is to just insist the branches be at the
> top level, rather than recursing down; we'd then recommend to organize
> them as projectA--main or projectA.main or projectA-main.

Meep. Eep. Thats very arch like, and wont solve the problem: you need to
read .bzr/branch/format for every branch and
check .bzr/branch/repository does not exist regardless, unless you also
disable nesting of repositories.

Really, do deal with latency here, solve the root cause :). Either make
the branch list opaque to the user, not a real directory on disk, so
that a system maintained compact list of branches is doable, or
eliminate the multiplying latency cost via async requests or a smart
server. Also note that WEBDAV should be able to give you all the
candidates in one request, and then one batch of requests should be able
to veryify the formats are known.

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/20060510/0d23636e/attachment.pgp