Re: [PATCH 24/34] checkout: reject if the branch is already checked out elsewhere

From: Junio C Hamano <gitster@pobox.com>
Date: Dec 3, 2014, 15:54 UTC
Message-ID: <xmqqegsgloeq.fsf@gitster.dls.corp.google.com>
In-Reply-To: <CACsJy8DLd4biEHo+FWYLqc4HKezHfv1ymWSQ8kLHLTQv=YKc3A@mail.gmail.com>

Duy Nguyen <pclouds@gmail.com> writes:

Show 14 quoted lines

> On Wed, Dec 3, 2014 at 12:30 AM, Junio C Hamano <gitster@pobox.com> wrote:
>>> I do like "read-only" ref concept where we can keep ref name
>>> (especially tags) in HEAD until the next commit. But it didn't go
>>> anywhere
>>
>> Remind me.  That sounds somewhat interesting.
>
> Couldn't find anything in my mail archive. But the idea is simple:
>
>  - we delay detaching HEAD until we need to update the associated ref
> at commit/reset time
>  - currently refs/tags/* are read-only. we generalize it to allow this
> 'read-only' attribute on some refs/heads/* as well. Because we can't
> really add attributes to refs, config var could be used.

It could be some annotation in HEAD instead, e.g.

    $ cat .git/HEAD
    ref: refs/heads/frotz
    options: read-only
    $ exit

and it may make more sense as this read-only-ness is closely tied to the state of HEAD, i.e. whatever operation that manipulates the HEAD, it needs to be very much aware of the read-only-ness, and it is much less likely to go out of sync, compared to a solution where you store this bit to anywhere else, e.g. config.