Recommended use of Bazaar for single-committer multiple-machine projects?

[Martin Pool]

On 13 Dec 2008, "Matthew D. Fuller" <fullermd at over-yonder.net> wrote:
> On Sat, Dec 13, 2008 at 01:04:16AM +0200 I heard the voice of
> Marius Kruger, and lo! it spake thus:
> > 2008/12/13 Martin Pool <mbp at canonical.com>
> >> I was just wondering (in the light of Mary's bug reports) whether
> >> we should in fact remove the --local commit feature.  It was a good
> >> experiment, and I think checkouts are really useful, but local
> >> commits not so much.
> >
> > Then you'd also have to remove unbind (because you can unbind,
> > commit, bind)
> 
> I wouldn't think so.  The reasoning for removing --local isn't that
> it's "broken" (it has breakage, but that breakage is fixable), but
> that it's a demonstrated source of confusion and trouble for users.
> I've not noticed bind/unbind to be so (IMO because they're never
> suggested as "Hey, you should try this...").
> 
> Even if it's _technically_ the same thing, it doesn't seem to carry
> the same traps socially.  I think that's because while, in a checkout
> sense, either way is "breaking the model" (you're taking the thing
> supposed to be in sync, and de-syncing it), --local is doing it sort
> of implicitly, while unbind is explicit.  While --local is a sort of
> weird "checkout" platypus, unbind is at least a straightforward "bound
> branch" thing.

Yes and no.

If we want to stop recommending to people that they should use --local,
we can of course do that with no changes.

I was talking about going a bit further than this and removing the
feature in the code too.  Though every bug in it is fixable, I'd
question whether fixing them one by one is really the best way to do it,
considering the complexity of the result for users and developers, and
and that there are other things we could be doing instead.

Matthew is right that if we removed this, we'd have to remove the whole
bound-branch feature.  The essential problem here is that rather than
just having a workingtree basis version and a branch tip version, we
also have a bound-branch latest version, and it can be out of date.
If that variable exists, it's unavoidable that it'll vary.

-- 
Martin      <http://launchpad.net/~mbp>