[RFC] bzr revert -r X..Y / -c X behavior

[Aaron Bentley]

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

John Arbash Meinel wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> So at the moment I'm quite happy with "bzr revert" and "bzr revert -r X"
> behavior. I was thinking, though, that we keep getting user requests for
> how to do "bzr merge . -r X..X-1" and thought that the command might be
> well handled by "bzr revert -c X".

I would really rather not make revert start doing merges.  That makes
our command set harder to understand.  Right now, people have a hope of
understanding what revert does, and adding modes makes it harder.  These
requests come from people who don't understand what revert does, and I
don't think they should spoil it for those who do.

> Considering that 'bzr revert -r -2' should do the same thing as 'bzr
> merge . -r -1..-2'.

*roughly speaking*.  Revert's not implemented on top of merge, and there
are good reasons for that.

> Probably the main issue is that we have a shorthand for cherrypicking
> merges (-c X) but not for revert cherrypicking (-r X..X-1, or -r
> X..before:X). We could possible do it with "bzr merge -C X", rather than
> on 'revert'.

I think that would be a better approach.

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

iEYEARECAAYFAknD21YACgkQ0F+nu1YWqI3mEQCfe3ZAhVUW6CVYIXR621objC+y
RYcAn332cNrI6uioUjjml33Y2Yv369Wc
=cKul
-----END PGP SIGNATURE-----