API change: juju subcommands will wait for connection to Zookeeper

[William Reade]

On Wed, 2011-11-16 at 15:35 -0700, Jim Baker wrote:
> One alternative was to put the waiting in the bootstrap subcommand, but
> this approach feels much cleaner. All subcommands that depend on ZK will
> simply be gated in this fashion.

Yep -- as I recall the motivation was that juju commands should block
until they've read/written all the state they're responsible for. So
bootstrap exits fast, because it does all it can quite quickly; the next
command may block for a while because it's waiting for state to exist,
but once it's performed its interactions it too can exit.

+1.