From: Junio C Hamano <gitster@pobox.com>
Date: Thu, 20 Aug 2015 15:49:24 GMT
Subject: Re: Why does 'git branch -d' act differntly after clone?
Message-ID: <xmqqa8tm7yez.fsf@gitster.dls.corp.google.com>
In-Reply-To: <CALwJaCzYgDKdcT7H7=5x0fnNRv4MsrtjJoHJcB9rJyh3dU--xQ@mail.gmail.com>

Konstantin Hollerith <khollerith@gmail.com> writes:

> Why does git branch -d acts differently after a clone?

In the former case, that branch is the only thing that knows about
the commits near the tip of it.  Immediately after you clone that
repository, you have a copy of that branch as a remote-tracking
branch for origin/neuerTestBranch.  You didn't say how you got a
local copy of it in there, but I'd imagine that in the clone you did
something like:

    $ git checkout neuerTestBranch

which is a short-hand for

    $ git checkout -b neuerTestBranch origin/neuerTestBranch

and then went back to a branch other than that branch when you asked
the branch to be removed, e.g.

    $ git checkout master
    $ git branch -d neuerTestBranch

without building any further history on neuerTestBranch.

Now, your local neuerTestBranch is a mere unmodified copy of the
remote-tracking branch origin/neuerTestBranch; the commits on the
former will not become unreachable if you remove neuerTestBranch,
because they are all reachable via origin/neuerTestBranch.  And that
is why the deletion of it has less severe potential of information
lossage, compared to the original case where you attempt to remove
the only copy.