[Kde-scm-interest] Have we arrived to a dead end?

[Oswald Buddenhagen]

On Sat, Feb 13, 2010 at 10:30:16AM +0100, Thiago Macieira wrote:
> Em Sexta-feira 12. Fevereiro 2010, às 23.59.43, Oswald Buddenhagen escreveu:
> > On Fri, Feb 12, 2010 at 11:19:29PM +0100, Thiago Macieira wrote:
> > > Atomicity and multiple repositories is impossible. Forget it.
> > 
> > not if the meta-push would be able to lock its submodules. but i've
> > already pondered that in another mail ...
> 
> Remember that there's no atomicity if you push two branches to the
> same repository..
> 
> One of them could fail and the other be updated.
> 
that might be the status quo, but i don't buy into any "this is
unsolvable" kind of argument.

the multi-branch scenario could be easily fixed by having update-ref
accept multiple pairs and employ internal locks. but this is not
relevant for this discussion.

the multi-repo pseudo-atomic push wouldn't be particularly hard to
implement either. the worst what could happen would be loss of atomicity
if the process was interrupted in the middle of updating multiple refs.
while that's clearly suboptimal, it isn't *really* a show stopper in that
context.
and if one wanted to go really overboard, one could implement a proper
transactional system in which the repos coordinate with each other
directly. i mean, distributed databases isn't really a completely new
research topic.