[MERGE] refactor of find_previous_heads from inventory to repository

[Aaron Bentley]

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

Robert Collins wrote:
> Also, I had thought that you'd be splitting the method into two: gather
> the revision_ids to find heads amongst from the inventory (which is
> inventory specific code), and secondly find the heads for (file_id,
> candidate_revisions).

Seems like it would also be nice to implement find_heads on Graph?  It
would also be nice to avoid get_ancestry, since it scales with the size
of ancestry.

Aaron
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGhJVV0F+nu1YWqI0RAhRJAJ0SRRSk6WPLUg0Wz7Crb+GuWsY7xwCeJe5Y
ee/TJ1suYjmxuf7WH5HyOPQ=
=Hpqw
-----END PGP SIGNATURE-----