Next Steps for hpss, etc

[John Arbash Meinel]

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

On 11/19/2010 8:27 AM, James Westby wrote:
> On Fri, 19 Nov 2010 14:50:37 +1300, Robert Collins <robert.collins at canonical.com> wrote:
>> It would be fantastic. I don't understand why this would let you use
>> stacked branches: AIUI you don't push the committed content anywhere,
>> and its discarded after the build. Wouldn't it be simpler to just not
>> commit?
> 
> Yes, it would.
> 
> Please contribute code to do this, or find someone else who can do
> it. It was beyond me, so I took the simple solution.
> 
> The issue is dealing with subsequent merges when you have pending merges
> and a modified tree.
> 

I thought 'bzr merge' took into account pending merges when finding the
new common ancestry. However, I honestly think it is *cleaner* from the
transformation perspective to split these out. It certainly is easier if
this wasn't just getting thrown away, since you could sort out what
happened where.

John
=:->

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (Cygwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAkzmneEACgkQJdeBCYSNAAOwwwCg2gW3zED+Ke7i//2ZrQLk/B44
fUUAoLU97k2ZsEWCdHuM4tTF/bVhxkUn
=i/Zo
-----END PGP SIGNATURE-----