[MERGE] handle null revision properly for LCA

[Aaron Bentley]

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

Aaron Bentley wrote:
> The null revision is root node of all revision graphs for all imaginable
> projects.  Its tree is the empty tree.  We have historically referred to
> it using None, but this means that we cannot use None in other contexts.

To emphasize this, the null revision is part of our model, not an API
hack.  If it is hacky, our model is hacky.  But has been in our model
since the baz 1.0 sprint.

We do not store the null revision-- we special-case it.  But calling it
a pseudo-revision suggests it's not real.  In terms of our model, it is
real.  I would much rather store it than try to remove it from our model.

>> Specifically having NULL_REVISION as a parent returned in queries means
>> that some revision ids returned cannot be accessed - theres no 'commit
>> date' for NULL_REVISION (and there shouldn't be).

There is a tree, however.  And I think that it would not be such a
terrible thing if there was an empty revision object associated with the
null revision.

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

iD8DBQFGeFk30F+nu1YWqI0RApoEAJ9i09+6ZhqhGxismdV1Cn4jO3ISJwCfW5VD
0QfRQSjaA/hxLgftzZBkNGU=
=9tTJ
-----END PGP SIGNATURE-----