the inventory must be updated as merge proceeds, not at the end

[Aaron Bentley]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Denys Duchier wrote:
> That sounds good.  It also sounds like what you are planning is going
> to make the work I have been putting into merge obsolete, so I think I
> should stop ;-)

Well, it's always a tough call, there, how much effort to put into
maintenance vs a rewrite.  But since you felt you couldn't really
comment further without seeing an implementation, I've started one.

> On the subject of abstractions, I'll show you mine (à charge de
> revanche) in case it helps.  I am by no means an expert pythonista,
> but I think I have manage to translate the gist of the design ideas I
> have mentioned previously.

Cool.  I'll have a look at that.  Here's mine:
http://code.albertanyone.com/bzr/bzr.ttransform/

I'm focusing on detecting as many possible conflicts as I can think of,
so it only handles creating new files, so far.  It's probably a bit
object-poor, in that it contains several maps of 'trans_id -> value',
where the set of trans_ids is heavily overlapping.  Those could probably
turn into FileEntry or something.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDvwFn0F+nu1YWqI0RAtKbAJ9DpulSvd6b5N/xWvDOtZwtz1NnEwCdEuKt
E3XkGknAXkGNNy3885SMGl8=
=8mUQ
-----END PGP SIGNATURE-----