[MERGE/RFC] Odd processing during BzrDir.sprout()

[Andrew Bennetts]

John Arbash Meinel wrote:
[...]
> Looking more closely, I think we need to address some stuff in BzrDir.sprout().
> 
> I see it doing:
>             source_branch = self.open_branch()
>             source_repository = source_branch.repository
> 
> But I never see it *locking* those objects. Which means it isn't caching any
> information between calls.

Also, I just noticed that one of the first things BzrDir.sprout does is call
self.cloning_metadir...  which opens a branch (or repository if that fails), and
a working tree, and then discards those objects.  So BzrDir.sprout inevitably
opens the source branch multiple times.  If the source branch is remote, that's
a fair bit of wasted effort.

So, lots of scope for improvement in this area.

-Andrew.