Weird failure of a merge to local branch

[Stephen J. Turnbull]

Aaron Bentley writes:

 > > That's true, but only until the next "bzr merge --pull" from the
 > > parent branch.  Then the two branches are again on the same revision.
 > 
 > Yes, because that's pulling, not merging.

Both in English and in all other VCSes "pull" merges, rather than
syncing.  (The fact that "push" is invariably restricted to a
fast-forward by default is a child-proof cap; it's not part of the
basic semantics.)  It's a real shame you guys didn't name "bzr sync"
as "bzr sync", and make it bidirectional.  Is it too late to at least
add that command as a standard alias?

I would expect that you're going to have more trouble with naming "bzr
sync --outgoing" as "bzr push" and "bzr sync --incoming" as "bzr pull"
when you get real colocated branches.  Note that in git there is
always a mirror branch (the so-called tracking branch).  This makes a
lot of sense (you have to fetch those commits anyway and they're part
of the DAG), and it makes a sync-from-remote operation completely
redundant.  "bzr pull" is a YAGNI.

Branches in bzr are more than just a name for a head, AIUI, but there
should be a way to handle this.