Checkouts vs bound branches

[Matthew D. Fuller]

On Mon, Jun 29, 2009 at 09:59:23AM +1000 I heard the voice of
Ian Clatworthy, and lo! it spake thus:
> 
> Right. See http://bazaar-vcs.org/DraftSpecs/SimpleCheckouts.

I for one continue to disagree.  And I think that spec demonstrates
some of the current problems by falling into them.

-------------
If you think a lightweight checkout is a heavyweight checkout with
zero history cached (like the spec author initially expected), you're
missing a key semantic difference:

    * lightweight checkout = tree & a reference to a branch
    * heavyweight checkout = tree+branch & a bind to a second branch.
-------------

That is indeed a problem, but the problem is that that "key semantic
difference" *IS* a key semantic difference of the implementation and
thought of as such, rather than as an unfortunate leak of an
implementation detail that was never fixed.

Checkouts and bound branches are two different beasts, and it's highly
unfortunate that they're similar enough that they were implemented as
if they were the same.  The solution to that isn't to declare that the
variant (heavy) of checkouts that is implemented as if it were a
synonym for bound branch no longer exists, it's to stop treating them
as synonyms and fix the places that leak.  I'm not interested in bound
branches, but I have plenty of use for checkouts that have a local
cache.


-- 
Matthew Fuller     (MF4839)   |  fullermd at over-yonder.net
Systems/Network Administrator |  http://www.over-yonder.net/~fullermd/
           On the Internet, nobody can hear you scream.