On Fri, Apr 15, 2016 at 10:36 PM, Junio C Hamano [off-list ref] wrote:
Reto Hablützel [off-list ref] writes:
quoted
the checkout command prevents me from checking out a branch in the
current worktree if it is already checked out in another worktree.
However, if I rebase the branch in the current worktree onto the
branch in the other worktree, I end up in a situation where the same
branch is checked out twice in the two worktrees.
I agree that any end-user facing subcommand like "git rebase", even
if it is not "git checkout", should refuse to work on and update a
branch that is checked out elsewhere. Otherwise it will end up
causing confusion.
I agree. I suppose we need same treatment for git-push? A push can be
rejected if the pushed ref is being checked out. Suppose HEAD is in
the middle of a rebase (or am), it fails to detect ref name (and thus
checkout state) the same way here and we definitely do not want "git
rebase" to simply overwrite whatever is updated by the push request.
--
Duy