[GSoC] Partial Cloning

[Rafael Goncalves Martins]

On 16:51 Sat 03 Apr     , Dirkjan Ochtman wrote:
> On Sat, Apr 3, 2010 at 05:13, Rafael Goncalves Martins
> <rafael at rafaelmartins.eng.br> wrote:
> > - Nallow clones
>
> Not sure if this is a typo: it's called 'narrow'. :)

Wow... My mistake! :)

> >    I think that the best approach here is implement the clonning of subtrees,
> >    storing the absolute path for the root of the subtree somewhere in .hg, and
> >    making Mercurial look for the subtree as part of a full "virtual" main tree,
> >    but ignoring the differences between the "virtual" tree and the files in the
> >    directories outside the subtree. This implementation should be hard, but it's
> >    the easiest way, IMO. I can be completelly wrong though.
>
> Are we agreed on keeping the whole manifest?

As I'm not following the whole discussion since the begin, I'm not sure. So
it's only my initial idea. Would be great if you or someone else can point me
to the previous design decisions.

> And what are you going to about (allow for) hg merge?

My initial idea is teach mercurial to only merge files inside the subtree, or
at least blocks the merge if some file outside the subtree is confliting.

Best Regards,

--
Rafael Goncalves Martins
http://rafaelmartins.eng.br/

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20100403/ae0bc8b4/attachment.pgp>