[MERGE] deprecated EmptyTree

[John Arbash Meinel]

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

Martin Pool wrote:
> On 21 Jul 2006, Robert Collins <robertc at robertcollins.net> wrote:
> 
>> I'd like to confirm - I think what you are saying is:
>> 'it sounds plausible to change the id of the root node between EmptyTree
>> and first commit rather than adding a root between EmptyTree and first
>> commmit'
> 
> That's what I was saying.  But maybe someone else has a reason why it's
> worse that way.
> 

Well, it is a little unclean for EmptyTree to have a root node. Since it
is theoretically empty.

Beyond that, once we get real root ids, the benefit pretty much
disappears. Robert wants it so that a tree which is empty has no
differences from EmptyTree. (Since right now all tree roots are ROOT_ID).

If there are no other reasons, then I would be against giving EmptyTree
a root id. There may be more theoretical reasons, like we want every
tree to have a root, etc. I don't think having root change its id is any
better than an explicit adding of the root id, but it isn't specifically
worse.

I'm just concerned that's Roberts request is just pushing the bugs
further down. Because once we have real roots, then even an empty
working tree won't match EmptyTree.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFEwM/TJdeBCYSNAAMRAi/vAKDYrjARrNS5TCrmFT6IpcEBRD4dCQCfelLq
BBvbmUozg/WwveaiX11nc6o=
=gd9J
-----END PGP SIGNATURE-----