bzr wouldn't update checkout

[Neil Martinsen-Burrell]

On 2009-09-15 10:04 , Matthew D. Fuller wrote:
> On Tue, Sep 15, 2009 at 09:47:33AM -0500 I heard the voice of
> Neil Martinsen-Burrell, and lo! it spake thus:
>>
>> Bound branches are necessarily a special case.
>
> So special bzr doesn't have them.  It also doesn't have heavy
> checkouts.  Instead it has a bastard mix.  This is the real problem,
> and it's terribly exacerbated IMAO by repeated injection[0] of bound
> branch terminology and concepts into discussions and onto people
> wanting checkouts.

I wholeheartedly agree.  The presence of a second branch in a 
heavyweight checkout (is that what we're supposed to call these bastard 
children until after 2.0?) is indeed the root cause of much confusion. 
I am looking forward to the day when there are two separate concepts: 
checkouts (always lightweight or at best with a local *cache*) and bound 
branches (full local branches that force a commit to another branch on 
every commit as a time/energy-saver vs. commit/push)

>> Another way to look at this is that ``bzr update`` in a bound branch
>> does more than just updating my working tree to match the remote
>> branch.
>
> I would argue that it shouldn't.  But that's unfixable without
> breaking checkouts until the cases are split, and I think checkouts
> are a much more important capability than bound branches.

We can disagree about the relative importance of bound branches vs. 
checkouts.  Personally, I find bound branches to be one of the most 
attractive feature of Bazaar and DVCS in general.  Thankfully, neither 
of these features will be going away and we can both be satisfied (see 
http://bazaar-vcs.org/DraftSpecs/SimpleCheckouts).

-Neil