Re: [PATCH] refspec: make @ a valid refspec

From: Felipe Contreras <felipe.contreras@gmail.com>
Date: Nov 25, 2020, 00:09 UTC
Message-ID: <CAMP44s3f_cO61YVjvwsiswKQRESznOFn7KTshgv06a=zVRPLug@mail.gmail.com>
In-Reply-To: <xmqqo8jmgyrd.fsf@gitster.c.googlers.com>

On Tue, Nov 24, 2020 at 4:45 PM Junio C Hamano <gitster@pobox.com> wrote:

Show 13 quoted lines

>
> Felipe Contreras <felipe.contreras@gmail.com> writes:
>
> >> I too find that "@ is a shortcut for HEAD" looks ugly both at the UI
> >> level and at the implementation level [*1*], but as long as we have
> >> it, it is good to try to be consistent and allow "@" everywhere
> >> where one would write "HEAD" in places where it is syntacitically
> >> infeasible---at least we would be consistently ugly that way ;-).
> >
> > Beauty is in the eye of the beholder.
>
> Using "@" is rather "illogical" than "ugly", and at that point it is
> not so subjective.

I disagree.

> @-mark leads a suffix that applies some "magic" to what comes before
> it (e.g. next@{1}, maint@{2.weeks.ago}, and master@{-1}).  Making @
> a synonym for HEAD means '@' sometimes means a ref and most of the
> time means the introducer of magic that applies to a ref.

All this was discussed back then [1].

In my mind "@" means the same as "@{0}", and "@{0}~0^0". So I don't see any inconsistency.

Your mind may be different.

> Worse yet, @{4} does not refer to HEAD@{4} but refers to the 4-th
> previous commit the current branch pointed at, so a mnemonic for the
> end user to remember the distinction between the two is that a bare
> "@" is different from HEAD, which is a total opposite X-<.

Once again; each user is different. Personally I never use these shortcuts precisely because I'm not entirely sure what I want to tell git. I just use "git reflog" and manually pick the commit id I want.

I suspect most users don't use this notation either, so even if there's an inconsistency (which I'm not entirely sure the @ shortcut has anything to do with it), it would affect few users.

> This is all water under the bridge, though ;-)

Indeed.

And I take it we can agree that it's better to have instructions like:

  git push -u origin @

Rather than:

  git push -u origin master

Regardless of what some users think of "@", it's less contentious than "master".

Cheers.

[1] https://lore.kernel.org/git/1367264106-2351-1-git-send-email-felipe.contreras@gmail.com/

-- 
Felipe Contreras