Changesets and multiple ancestors

[Aaron Bentley]

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

John A Meinel wrote:
> Aaron Bentley wrote:
> 
>>
>> The behavior isn't as nice because when it's time to merge K, because
>> the best base is G, and that's not available.
> 
> Except K has to have G because it branched from it.
> Your local tree doesn't have it. But someone among the 2 does.
> And if you generated a changeset, you've already done the work, and I
> can just merge what you say to.

If I send you a changeset for K, it does not provide a way for you to
produce G.  It also doesn't provide a way for you to avoid using G.
Remember that the changeset for K is equivalent to a tarball of the K
working tree, plus appropriate metadata-- it's just an optimization.
Unlike Arch changesets, bzr changesets are not supposed to be applied
using inexact patching.

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

iD8DBQFCwYWH0F+nu1YWqI0RAkEHAJ9qWzQJUaEP6hmoNGbl1kTBjbJjaACdEtms
HjJ2fbaoOPsUg1mbSh/KN9s=
=gcDU
-----END PGP SIGNATURE-----