[TERMINOLOGY] rename "changesets" to "revision bundles"

[Aaron Bentley]

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

Wayne Davison wrote:
> On Fri, May 26, 2006 at 04:33:56PM -0400, Aaron Bentley wrote:
> 
>>I'd like to propose we rename "changesets" "revision bundles".
> 
> 
> What about the term "revision patch"?  I like the idea of leveraging
> the term patch, which is well understood to mean something that can be
> used to instantiate a set of changes.

I think that implies some things that are incorrect
- - that it is produced by "bzr diff"
- - that there is a "bzr patch" command to apply "revision patches" (there
is a "patch" command in bzrtools, but it's for plain patches.)
- - that it contains a single revision
- - that it has the same failure modes as patch (that is, if both sides
have applied a patch, it will produce .rej hunks)-- in fact, since it's
a revision-transmission mechanism, it is as good as the merge algorithm,
which is be
- - that it lacks metadata about renames, execute changes, symlink
targets, etc.

> As for what the sub-command should be to generate a revision patch, I'd
> suggest rdiff, and perhaps revision-diff as a longer alias.  

I think this has the wrong connotations, because it suggests that the
"revision patch" contains just the diff between the base and target
revisions, rather than a complete set of data for installing all the
ancestors of target which are not ancestors of base.

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

iD8DBQFEewSD0F+nu1YWqI0RAoV+AJ93N+mEp6ZZDGrDcXuPhJju1BmDeACcCC09
84VWvS0tGkSwGE2VzlncjPc=
=BsHe
-----END PGP SIGNATURE-----