[MERGE REQUEST] New merge base picker

[Aaron Bentley]

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

Hi all,

I've implemented a new algorithm to pick a merge base.  The old one had
poor scalability, and it could encounter situations in which there was
more than one answer.  This new algorithm addresses both problems.

It can be found at the usual place:
http://panoramicfeedback.com/opensource/bzr.24

In the common case, this algorithm behaves like the old one.  In some
corner cases, it behaves differently.

The old algorithm found the shortest path from THIS and OTHER to a
common ancestor.  This could produce contradictory results, since the
closest node to THIS might not be the closest node to OTHER.

The new algorithm finds the longest path from the graph root to a common
ancestor.  This imposes a new requirement that the graph have a single
root.  I do not think this is a problem, since all revisions ultimately
derive from the NULL revision.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDI3+x0F+nu1YWqI0RAjt6AJ44nNICiRItJSuJM8DY4262RssTpwCfZp1H
Qkwdg/m95paS+CaJO3HaZME=
=s/a/
-----END PGP SIGNATURE-----