[Gsoc] Partial clones

[Peter Arrenbrecht]

On Sat, Jul 12, 2008 at 5:15 PM, Frédéric <fred.rec at gmail.com> wrote:
> Hi,
>
> I realize that I did a very poor communication on my project, so lets
> correct that.

Splendid!

> In my application I estimated that the first milestone (a working
> implementation of narrow clones) would have been reached just before
> the mid term evaluation. However the progress was slower than
> expected.  Particularly because I overestimated the time I was able to
> devote during the firsts weeks of June and especially because I
> underestimated the complexity of the task.
> Because of Mercurial's structure, implementing the narrow partial
> clones needs a kind of shallow cloning, what requires a deep
> understanding of the revlogs.
>
> I tried various methods, but most of them worked only on particular cases.
>
> My current approach is based on Parren's shallow clones and a good
> part of the work is implemented.
> I have to deal with some manifest related problems and to aggregate
> all the small pieces of code I've ever done.
> So I hope that a working implementation will be ready within the next week.

I'm intrigued. I would not myself have thought them to be so closely
related. Is it possible to take a look at your work somewhere?

Cheers,
-parren